Development of the MEDOS/HIA DeltaStream Extracorporeal Rotary Blood Pump

Artificial Organs 25(5):358–365, Blackwell Science, Inc. © 2001 International Society for Artificial Organs

Development of the MEDOS/HIA DeltaStream Extracorporeal Rotary Blood Pump *Christof Go¨bel, *Arash Arvand, †Rolf Eilers, †Oliver Marseille, †Christoph Bals, ‡Bart Meyns, ‡Willem Flameng, *Gu¨nter Rau, and *Helmut Reul *Helmholtz Institute for Biomedical Engineering, Aachen University of Technology, Aachen; †Medos Medizintechnik AG, Stolberg, Germany; and ‡Department of Cardiac Surgery, University of Leuven, Leuven, Belgium

Abstract: The DeltaStream blood pump has been developed for extracorporeal circulation with one focus on potential integration into simplified bypass systems (SBS). Its small size and an embedded electric motor are the basic pump properties. A variation of the impeller design has been performed to optimize hydraulic and hematologic characteristics. A simple impeller design was developed which allows flow and pressure generation for cardiopulmonary bypass applications. The option of a pulsatile flow mode for ventricular assist device applications also was demonstrated in vitro. Impeller washout holes were implemented to improve nonthrombogenicity. The pump was investigated for potential thermal hazards for blood caused by the integrated electric motor. It could be demonstrated that there is no thermal risk associated with this

design. Durability tests were performed to assess the lifetime of the pump especially with regard to the incorporated polymeric seal. Seal lifetimes of up to 28 days were achieved using different blood substitutes. In animal tests using either the pump as a single device or in an SBS setup, biocompatibility, low hemolysis, and nonthrombogenicity were demonstrated. In summary, the DeltaStream pump shows great potential for different extracorporeal perfusion applications. Besides heart-lung machine and SBS applications, ventricular assist and extracorporeal membrane oxygenation up to several days also appear promising as potential applications. Key Words: Rotary blood pump—Simplified bypass system—Cardiopulmonary bypass—Pulsatile flow—Impeller design—Thrombogenicity—Durability.

Techniques of extracorporeal circulation (ECC) of cardiac patients have become diversified in recent years. While the conventional heart-lung machine was the standard of mechanical perfusion over many decades using either roller or centrifugal pumps, today new techniques and systems have penetrated the clinical field or are under development. Examples are extracorporeal right heart pumps to maintain lung perfusion intraoperatively (1) or techniques to assist only the left ventricle during surgery when autonomous right heart function is sufficient (2). Some devices assist the ventricles intracorporeally positioned within the aorta or pulmonary artery (3,4). Further needs for mechanical perfusion exist beyond cardiac surgery. Patients with cardiopulmonary

bypass (CPB) weaning problems, acute or chronic heart failure, and even lung diseases require mechanical circulatory support (5–9). In many of these cases, rotary pumps meet the requirements of ECC in terms of hydraulic power, lifetime, nonthrombogenicity, and cost. However, there also are certain limitations with rotary pumps regarding durability, thrombogenicity, and pulsatility (10–14). Many extracorporeal rotary pumps have integrated seals which limit their lifetime. Pumps without seals, using blood immersed bearings and magnetic coupling, are basically limited by an inherent risk of thrombus formation only. However, in longer term ventricular assist device (VAD) applications up to several days, centrifugal pumps offer an attractive alternative to pneumatic or electromechanical VAD systems due to their low price even if they are officially approved for only 6 h. An approach which has been followed at the Helmholtz Institute is the miniaturization of circulatory assist devices by means of a small rotary pump.

Received January 2001. Presented in part at the 8th Congress of the International Society for Rotary Blood Pumps, held September 6–9, 2000, in Aachen, Germany. Address correspondence and reprint requests to Dr. Christof Go¨bel, Helmholtz Institute for Biomedical Engineering, Pauwelsstrasse 20, D-52074 Aachen, Germany.

358

EXTRACORPOREAL ROTARY BLOOD FLOW An example of such miniaturization is a simplified bypass system (SBS), which is a closed loop heartlung machine that is reduced to its basic components, the pump and the oxygenator (Fig. 1). Such an SBS can be used on the operating table in the case of heart surgery or close to the patient in emergency cases during transport. Avoiding a venous reservoir, cutting down tubing length, and reducing extracorporeal blood volume are considered the major virtues of such a system. Secondary effects of short tubing are the reduction of foreign surfaces as well as of the extracorporeal pressure difference. The associated blood trauma would be decreased by a reduction of pressure difference and pump speed. Another consequence of lower extracorporeal blood volume is a decreased need for blood transfusions with their associated costs. This approach combines the safe and highly accepted standard of CPB with a reduction in size, extracorporeal blood volume, artificial surfaces, and blood damage. The development of the DeltaStream pump was initiated about 3 years ago at the Helmholtz Institute with the main focus on SBS as described previously. However, the pump also was designed with sufficient hydrodynamic power for standard CPB. A further idea was to use the DeltaStream in extracorporeal life support (ECLS) applications (7,9) or as a medium-term VAD system up to several days. The DeltaStream is mechanically independent from a driving console in contrast to the classical centrifugal pump since the driving motor has been incorporated into the pump. Thus, the pump has only an electrical connection to the driving console which works as the power supply and the control unit. This design allows positioning of the device close to the patient in the operating theater or using

FIG. 1. The schematic drawing is of an SBS consisting of a pump and an oxygenator.

359

it as a VAD system on the patient’s belly, similar to pneumatically driven VAD systems. MATERIALS AND METHODS The basic design idea of the DeltaStream is characterized by the following features. It has an embedded long cylindrical electromotor. This motor is surrounded by an annular flow path, which provides sufficient motor cooling by blood under all possible operating conditions. The impeller is positioned between the pump inlet and the motor housing, and the motor and the impeller have basically the same diameter. A polymeric seal separates the blood chamber and the motor chamber. Figure 2 shows a computer-aided design (CAD) study of the DeltaStream lab type. With an integrated electric driving motor, a further challenge is present. Heat, dissipated by the motor, must not result in overheated surfaces on the blood side or at the outer housing. With a focus on application in SBS but keeping alternative perfusion techniques in mind, the resulting pump requirements were set as follows. The pump should be small, but hydraulic power should minimally meet an operating point of 400 mm Hg pressure difference at 7 L/min flow. Hemolysis should be comparable to the BP 80, and the design should have low thrombogenicity to allow applications such as VAD or in ECLS for at least 24 h. To obtain such a low thrombogenicity, the design should totally avoid low flow areas as they typically appear at the rear sides of the impeller. Therefore, suitable patterns have to be generated within the pump. When starting the development, we were not sure whether a realization of a small pump with low hemolysis properties in CPB applications was possible.

FIG. 2. A CAD study of the DeltaStream lab type is presented. Artif Organs, Vol. 25, No. 5, 2001

360

¨ BEL ET AL. C. GO

Looking at centrifugal pumps designed for CPB, it is obvious that their impellers have a rather large diameter that allow propelling blood at low rotational speeds. Even if the circumferential speed is considered the most important parameter, the realization of an impeller with a diameter of about 25 mm was a challenge. One major focus, therefore, was a variation in the impeller design and assessment of associated hydrodynamic and hemolytic properties. The realization of a durable mechanical seal and a suitable drive unit were further tasks. Basic impeller design The variables for the impeller design were a meridional shape of hub and vanes, number of vanes, wrap angle, vane shape, and angles of entrance and exit, respectively. An overview of the design modifications can be seen in Fig. 3. Three different meridional shapes were chosen; the main difference is the position of fluid entrance with regard to the vanes. This fluid entrance is stepped back with Type 2 and even further retracted with Type 3, i.e., the blades are shorter in their circumferential projection. One 2-vane design has been realized, using a cross section of Type 1. Two 3-vane and one 4-vane impellers also were designed. Impeller 2 has a cross section of Type 2, and Impellers 3 and 4 have a cross section of Type 3. Impellers 1 to 3 have curved vanes with their blade angles calculated on the basis of turbomachinery principles whereas Impeller 4 was designed under manufacturing aspects. In addition, the impellers were modified by introducing a shroud to minimize the internal fluid leakage caused by the axial pressure difference along the impeller.

Nonthrombogenicity measures With any centrifugal or mixed flow pump, the impeller rear side is a critical region in terms of washout and risk of thrombus formation. Addressing this aspect, Impeller 4 was supplied with additional washout holes. These holes assist washout of the crucial shaft seal site by hydraulically short-cutting pressure and suction on the suction side of the impeller. Fluid is sucked from the shaft-seal site and, after passing the holes, reenters the main flow region. These holes originate directly from the shaft so that there is no dead volume which could cause blood stagnation or persistent recirculation phenomena. Bore chamfers allow a smooth redirection of the blood back into the mainstream, avoiding flow separation. A hole with a diameter of 3 mm was chosen. The drive unit A brushless DC motor was chosen as the drive unit. This motor has a stainless steel casing and is embedded within the bloodstream. The DC motor has an electromechanical efficiency of up to 90%, which is in the range of the best motors available today. Efficiency may vary under certain hydraulic conditions and finally depends on actual motor speed and impeller loading, but heat dissipation is low under all possible hydraulic pump conditions. A motor with stationary coils and moving magnets was chosen since with this motor principle there is no air gap between the coils and blood. Blood works as a heat drain with a low heat transfer resistance between coils and blood. Thus, heat loading of the blood by the motor and storage of heat in motor components are not expected. The DeltaStream prototype A list of materials and dimensions of the DeltaStream is given in Table 1 for the lab type and production type pumps, respectively. Polymethylmethacrylate and thermoplastic polyurethane were used for lab type pumps since all parts were machined and both materials are easy to polish. Stainless steel 303 was used for easier machining of the lab types, but it has less corrosion resistance compared to 316L used in the production type pumps. The housing wall thickness was reduced in the production type pumps, which resulted in a smaller diameter and weight. EXPERIMENTAL STUDY

FIG. 3. Shown are cross-sectional views and 3-D CAD studies of the impeller design variations. Artif Organs, Vol. 25, No. 5, 2001

Hydraulic characteristics Hydraulic investigations were performed using a water glycerin mixture with a dynamic viscosity of 3.6 mPas as the model fluid. The main focus of these

EXTRACORPOREAL ROTARY BLOOD FLOW

361

TABLE 1. Materials and dimensions of the DeltaStream Lab type pump Pump housing Motor housing Impeller Pump length Pump diameter Priming volume Pump weight

Acrylate (PMMA) Stainless steel 303 Thermoplastic polyurethane 130 mm 45 mm 30 cc 350 g

Production type pump Polycarbonate Stainless steel 316L Polycarbonate 150 mm 40 mm 30 cc 280 g

PMMA: polymethylmethacrylate.

tests was to assess the effects of variations in vane number and geometry. The effect of impeller shrouding was investigated with Impellers 1 to 3. For Impeller 4, only a shrouded version was tested. The clearance from the impeller tip to the housing or the cover plate to the housing, respectively, also was varied. In addition to these tests, the first investigations in a simple pulsatile mode were performed within a circulatory mock loop, which simulates the basic elements such as resistance, fluid inertia, and aortic compliance. In this preliminary test, the pump was driven with a sinusoidal variation of the rotational speed. Impeller 1 with a cover plate was used for these tests. Thermal investigation As previously mentioned, an important aspect during pump development was the thermal loading of blood by the embedded electric motor. A copperconstantan thermoelement was fixed onto the motor surface to assess motor surface temperature during pump operation under extreme hydraulic pump conditions. The sensor position in the pump can be seen in Fig. 4, in which the motor itself is not shown for

FIG. 4. The CAD study shows the motor removed and an additional canal in which the temperature sensor was fixed onto the motor surface for thermal investigations.

better insight. A small canal was milled into the stainless steel cage in which the sensor was fixed onto the motor housing. Thermally conductive paste was used to ensure that the sensor easily adapts to motor temperature. The pump delivered fluid from a heated reservoir. Fluid temperature was kept constant at 37.5°C within the reservoir. The pump operating conditions were set to a ⌬P of 400 mm Hg at a flow of 6 L/min. Motor temperature was monitored during pump operation and after sudden pump stop. Durability tests Technically, pump durability mainly depends on the lifetime of the shaft seal under specific operating conditions. In clinical practice, additional limitations may result from eventual thrombus formation in the pump. The in vitro durability tests were focused on the investigation of seal lifetime using different blood substitutes such as water, saline, and hydroxyethylstarch (HES) plasma expander. Long-term in vitro tests with blood are not feasible due to blood handling problems. A special tester was designed to simplify assembly and to perform tests with a larger number in order to obtain statistical significance. Typical load characteristics for the seals are simulated in this tester. Up to 8 test units can be run simultaneously. Motor units are positioned in a special holder made of brass. The seal, fixed into a seal holder, is positioned between the test chamber and the motor holder. The test shafts, supported by 2 ball bearings, are coupled to the brushless DC motor. The drive units display motor speed and current, where the motor current is proportional to the shaft-seal friction. Thus, information about the seal status can be obtained via the motor current reading. Tests were performed under typical CPB conditions, such as a ⌬P of 350 mm Hg and a rotational speed of 8,000 rpm. In a pump, these conditions would correspond to a flow of about 5 L/min. Tests have been performed with a duration of up to 10 days. In vitro tests with porcine blood also were perArtif Organs, Vol. 25, No. 5, 2001

362

¨ BEL ET AL. C. GO

formed to test seal function in blood. Maximum test duration with blood was 10 h. Several seal geometries and material pairings were studied during pump development. Minimum seal and pump durability was defined to be 5 days. Using a safety factor of 2 for the in vitro tests, the test duration was set to 10 days for the final seal design. A total of 32 tests was performed with the final design with a duration of 10 days. Additional durability and reliability tests have been performed with fully assembled production type pumps.

Protocol 2 In the second protocol, the pump was used together with a MEDOS HiLite 7000 oxygenator. Six experiments were performed according to Protocol 2, in which early production type pumps of the DeltaStream were used. No loop was realized since oxygenated blood is returned to the animal. The extracorporeal circuit worked without a venous reservoir. The pump and oxygenator represented a closed loop SBS. Q=

Animal experiments Animal experiments, using sheep as a model, have been performed with lab types and early production type pumps of the DeltaStream. Sheep were chosen for the experimental animal model because their weight and hemodynamic behavior are similar to humans. The animals were treated in compliance with the Guide for the Care and Use of Laboratory Animals (NIH publication No. 85–23, Revised 1985). The goals of animal testing were to assess pump handling, overall biocompatibility, blood cell trauma, and thrombogenic properties. Experiments were performed either using the DeltaStream alone or with an oxygenator as a closed loop SBS. Suitable protocols were developed for both settings. Cannulas were inserted into the internal jugular vein and the common carotid artery. Blood gases were monitored during the experiment. Protocol 1 The DeltaStream was used without an oxygenator, and a certain loop was realized which allowed running the pump under defined conditions. One experiment was performed according to Protocol 1, in which a final lab type of the DeltaStream was used and an operating point of 6 L/min at 350 mm Hg was chosen. Since in this protocol venous blood is returned to the animal from the extracorporeal loop, this partial blood flow had to be kept low, in the range of 0.5 L/min, so that blood gases were not substantially deteriorated. The remaining 5.5 L/min was circulated within the loop. Two clamps were used to adjust the balance between the animal and the loop flow. The experimental duration was 8 h. Physiological and hematological data were taken during and 7 days after the experiment. The pumps were inspected for thrombus formation after the experiment. The animals were sacrificed on the seventh day after experiment and necropsy was performed to show potential signs of organ infarction, which could have been caused by thromboembolic events. Artif Organs, Vol. 25, No. 5, 2001

冉

冊

Wt共kg兲 + 0.4 × 2.4 50

(1)

Flow was adjusted according to animal weight based on Eq. 1, resulting in flows of 3 to 4 L/min and according ⌬P of 200 to 250 mm Hg across the pump. Experimental duration was 6 h. Inspection of the pump after the experiment, blood data recording, and necropsy on the seventh day after the experiment were performed as in Protocol 1. In all experiments, 300 IU/kganimal weight of heparin was administered initially. Later, 5,000 IU was added to keep the activated clotting time (ACT) above 200 s. RESULTS Hydraulic characteristics The hydraulic charts for different impellers are shown in Fig. 5. Sample graphs are shown for Impeller 1 without shroud, Impeller 3 with and without shroud, and for Impeller 4, which was only tested in a shrouded version. The speed range for all graphs is 5,000 to 10,000 rpm. The basic findings are summarized next.

FIG. 5. The graphs depict the hydraulic characteristics of different impeller design variations.

EXTRACORPOREAL ROTARY BLOOD FLOW Some differences were found in pressure flow characteristics between the different unshrouded impeller types. When comparing Impellers 1 and 3, the design of Impeller 3 obviously had a lower pressure generation. From all tested versions, Impeller 3 in the unshrouded version had the lowest pressure build-up. A notable effect on pressure build-up was found by using impeller shrouds, demonstrated with Impeller 3. At a rotational speed of 10,000 rpm and a flow of 6 L/min, a pressure of 340 mm Hg is generated by the unshrouded version whereas the shrouded version generates about 500 mm Hg under the same conditions. Comparing the hydraulic charts of shrouded Impellers 3 and 4, a high similarity can be seen in the low flow region. A remarkable finding is the behavior of Impeller 4 within the high flow region. The decline of the characteristic curves is much less with this impeller design, resulting in higher pressure generation at flows from 6 to 10 L/min. All curved vane designs, including Impeller 2 which is not shown here, reveal similar pressure flow relations when a shroud is used. Impeller 4 is outstanding possibly due to the rather centrifugal character of its blades. The charts shown in Fig. 5 were found for different clearances between the vane tip and the housing or the shroud and the housing, respectively. With unshrouded impellers, a clearance in the range of 0.2 to 0.3 mm had to be realized. Increasing this gap results in a dramatic decrease of pressure generation. This investigation was performed with Impellers 2 and 3. In contrast, shrouded impellers allow a clearance of 1.5 mm without impairment of hydraulic characteristics. Varying this clearance in steps of 0.3, 0.7, 1.1, 1.5, and 2.0 mm, an obvious decrease in pressure generation can be observed at 2.0 mm; all other clearance variations had only minor effects on the characteristic curves. Sample graphs are shown for clearances of 0.3 mm for unshrouded impellers and 1.5 mm for shrouded impellers. Further investigations of Impeller 4 with additionally integrated washout holes did not influence the hydraulic characteristics. Figure 6 shows the results from an early and very simple pulsatile test. As can be seen from the graphs, rotational speed varied from 3,000 to 9,500 rpm to generate a pressure fluctuation between 80 and 120 mm Hg within the circulatory mock loop. Peak flow reached 17 L/min. The pump inlet pressure varied between 5 and −45 mm Hg. Obviously, the pressure and flow curves have a sinus wave form due to the sinusoidal variation of rotational speed.

363

FIG. 6. Speed, flow, and pressure curves based on a simple pulsatile test are shown.

Thermal investigations Figure 7 shows the motor surface temperature under 2 conditions. First, temperature was recorded under a hydraulic pump load of 400 mm Hg pressure difference and 6 L/min flow as a worst case situation. A thermal equilibrium was found with a motor surface temperature of 38.1°C, which was 0.6°C above the reservoir temperature. Upon sudden pump stop, the maximum temperature increased to 38.4°C, which was 0.9°C above the reservoir temperature. This peak temperature was reached at 13 s after the pump stopped. The initial temperature of 38.1°C was reached again 10 s later. After an additional 60 s, the motor surface cooled down exactly to the reservoir level. The steps of 0.1°C within the graph represent the sensitivity of the data acquisition equipment. Durability testing Of the 32 tests with the final design, none failed within the defined period of 10 days, after which the tests were terminated. With the production type pumps, much longer lifetimes were obtained.

FIG. 7. Shown is the motor temperature during pump operation (at 400 mm Hg, 6 L/min) and after a sudden pump stop. Artif Organs, Vol. 25, No. 5, 2001

¨ BEL ET AL. C. GO

364

Animal experiments In the animal studies, the DeltaStream was easy to handle and showed excellent feasibility. Filling the pump and circuit and de-airing were easy to perform. The placement of venous cannulation was delicate in the SBS setup, where flows of about 3 L/min had to be achieved according to Eq. 1. Venous cannulas with 28 or 34 Fr size, respectively, were used. Larger cannulas could not be used due to the diameter limitation of the jugular vein. In some cases, initial suction occurred due to a limited venous blood supply, and the predetermined flow had to be reduced. Figure 8 shows free plasma hemoglobin over time for the performed tests. Hemolysis was baseline in all experiments with some variation over time. This was found for the pump alone as well as with the pump in an SBS. Recordings showed an ACT of 200 s and higher during the experiments. There was no thrombus formation throughout all experiments. Particularly, the crucial impeller rear side and washout holes were free of thrombi. No signs of organ infarction were found at necropsy. COMPUTATIONAL FLUID DYNAMICS In addition to the experimental investigations of pump properties, the design also was validated by computational fluid dynamics (CFD), for which the commercially available software package TASCFlow (AEA Technology, Otterfing, Germany) was used. A grid was generated for the entire flow domain with approximately 340,000 elements. Blood was regarded as Newtonian fluid with a viscosity of 3.6 mPas for the calculations. An example of the CFD results is shown in Fig. 9, in which a distribution of velocity vectors in a projection into the meridional cross section is shown. A major focus of the

FIG. 8. The graph represents the plasma free hemoglobin levels of 7 animal experiments. Artif Organs, Vol. 25, No. 5, 2001

FIG. 9. Shown are the projected velocity vectors in a meridional cross section based on a CFD simulation.

CFD studies was leakage flow through the washout holes of Impeller 4. Calculations revealed a flow between 0.5 and 1.4 L/min per hole depending on rotational speed and pressure difference. DISCUSSION AND CONCLUSIONS Impeller 4 in the shrouded version showed the best hydraulic characteristics with only minimum decrease of pressure build-up at higher flows. This may be due to the centrifugal character of its vanes. With this design, later injection molding was expected to be much easier compared to the curved vane designs. For these reasons, Impeller 4 was integrated into the final pump. Additional washout holes appear as an effective modification which contributes to pump washout and thus nonthrombogenicity without any measurable impairment of hydraulic characteristics. Only a slight increase in the impeller driving torque in the range of 10% was observed, which indicates that the impeller efficiency is decreased by this amount. This can be interpreted by a recirculation flow with associated hydrodynamic losses. Numerical calculations revealed a high flow and velocity at the impeller rear side and within the holes. Therefore, Impeller 4 with washout holes was chosen as a compromise between hydraulic efficiency and washout. A shrouded impeller was chosen for the final design to minimize the internal leakage flow and to allow the realization of a large clearance. Again, with a focus on manufacturing and assembly procedures, a larger gap appears as a superior solution since potential problems with manufacturing tolerances, e.g., central positioning of the impeller within the housing, are minimized. The final production type pump is shown in Fig. 10.

EXTRACORPOREAL ROTARY BLOOD FLOW

FIG. 10. The photograph is of the production type of the DeltaStream pump.

Thermal investigations of the DeltaStream did not reveal any risk of blood loading by dissipated heat. In a pump lab type, the design showed only a low increase in the motor temperature during high motor loading and after a sudden pump stop. A sudden pump stop during operation is generally regarded as the most crucial situation since energy which is stored in motor and pump components is released to the then stagnant blood. Temperature recordings were taken at a location between the motor surface and the casing. Therefore, in all situations, the actual surface temperatures exposed to the blood were even lower than the measured data. Seal fatigue tests using blood substitutes such as saline or HES showed the potential of the DeltaStream to be used in longer term applications. With seal lifetimes of 10 days and more, the pump may be very suitable for extrcorporeal membrane oxygenation or as a VAD system. The comparability of results obtained with blood substitutes is, of course, limited due to the complexity of blood with its cellular components, potential of fibrin deposition, and risk of thrombus formation at the shaft seal site. In animal experiments, overall pump biocompatibility, ease of handling, and low blood damage were demonstrated. No thrombus was found in the experiments, and low thrombogenicity also is expected in clinical situations. Of course, with ACT times of 200 s and higher, the significance of the findings is limited. In clinical routine, particularly in the postoperative phase, ACT times might be kept lower to control bleeding. Further studies at lower ACT times should be performed to aggravate test conditions and to obtain further information about thrombogenic properties of the pump.

365

With the presented results, the DeltaStream has shown great potential for different applications in ECC. The hydraulic capabilities are sufficient for applications in conventional CPB, and hematological results show low blood damage even under extreme hydraulic conditions. In addition, the pump design and mechanical independence from a driving console make operation in compact perfusion systems like an SBS very attractive. With the possibility to position the DeltaStream on or very close to the patient, it also can be an alternative to expensive pneumatic VAD systems. With the demonstrated potential to generate physiologic pulsatile flow, such an application of the DeltaStream appears even more attractive. REFERENCES 1. Reul H. Right heart support, percutaneous assist devices— What is new? Cardiovascular Engineering (in press). 2. Waldenberger FR, Haisjackl M, Holinski S, Lengsfeld M, Konertz W. Centrifugal pumps as left ventricular assist for coronary revascularization on a beating heart. Artif Organs 1998; 22:698–702. 3. Meyns B, Sergeant P, Siess T, Zietkiewicz M, Perek B, Nishida Y, Flameng W. Coronary artery bypass graft with biventricular microaxial pumps. Perfusion 1999;14:287–90. 4. Lo¨nn U, Peterze´n B, Carnstamm B, Casimir-Ahn H. Beating heart coronary surgery by an axial blood flow pump. Ann Thorac Surg 1999;67:99–104. 5. Noon GP, Ball JW, Short HD. BioMedicus centrifugal ventricular support for postcardiotomy cardiac failure: A review of 129 cases. Ann Thorac Surg 1996;61:291–5. 6. Curtis JJ, Walls JT, Schmaltz RA, Demmy TL, Wagner-Mann CC, McKenney CA. Use of centrifugal pumps for postcardiotomy ventricular failure: Technique and anticoagulation. Ann Thorac Surg 1996;61:296–300. 7. Bartlett RH, Roloff DW, Custer JR, Younger JG, Hirschl RB. Extracorporeal life support. The University of Michigan experience. JAMA 2000;283:904–8. 8. Lewandowski K, Roissant R, Pappert D, Gerlach H, Slama KJ, Weidemann H, Frey DJM, Hoffmann O, Keske U, Falke KJ. High survival rate in 122 ARDS patients managed according to a clinical algorithm including extracorporeal membrane oxygenation. Intensive Care Med 1997;23:819–35. 9. Kolla S, Awad SS, Rich PB, Schreiner RJ, Hirschl RB, Bartlett RH. Extracorporeal life support for 100 adult patients with severe respiratory failure. Ann Surg 1997;226:544–66. 10. Curtis JJ, Boley TM, Walls JT, Demmy TL, Schmaltz RA. Frequency of seal disruption with the Sarns centrifugal pump in postcardiotomy circulatory assist. Artif Organs 1994;18: 235–7. 11. Paul R, Marseille O, Hintze E, Huber L, Schima H, Reul H, Rau G. In vitro thrombogenicity testing of artificial organs. Int J Artif Organs 1998;21:548–52. 12. Tayamy E, Ohtsubo S, Nakazawa T, Takami Y, Niimi Y, Makinouchi K, Glueck J, Nose´ Y. In vitro thrombogenic evaluation of centrifugal pumps. Artif Organs 1997;21:418–20. 13. Go¨bel C, Eilers R, Reul H, Schwindke P, Jo¨rger M, Rau G. A new blood pump for cardiopulmonary bypass: The HiFlow centrifugal pump. Artif Organs 1997;21:841–5. 14. Yada I. Is nonpulsatile flow detrimental to patients? Artif Organs 1993;17:291–2.

Artif Organs, Vol. 25, No. 5, 2001