The causes of heart failure are numerous. Some diseases, such as idiopathic cardiomyopathy, develop over a long period of time, sometimes leading to heart transplantation, whereas others occur suddenly during acute myocardial infarction or after heart surgery. 1–3 In many instances, medical treatment fails, and mechanical support is the only possible treatment. Most of the support systems are extremely invasive, requiring major surgery, although a few are either percutaneous or only require minor surgery, such as the intra-aortic balloon pump, the Hemopump, and percutaneous cardiopulmonary bypass. 4–8
We are currently developing an axial flow pump that consists of a flexible catheter with a stiff pump head at the distal end and a drive unit at the proximal end. To make the pump head as small as possible during insertion, it consists of a foldable propeller and a foldable cage, the latter consisting of six filaments. The cage protects the wall of the vessel against the propeller and vice versa (Figure 1, A and B).
The catheter pump is intended to be inserted into the femoral artery in the groin and be placed in either the ascending or descending part of the aorta. There are theoretically four rationales behind the pump. 1) Because both the active hydraulic part (propeller) and cage are folded, the catheter is thin when inserted into the arterial system, but has the properties of a big propeller when deployed and activated. 2) If the propeller is placed in the aorta, either in the ascending or descending part, lower pressure in the aortic arch and above the aortic valve will occur, resulting in afterload reduction for the left ventricle. 3) Distal to the propeller, higher pressure is generated, facilitating the perfusion of the lower parts of the body. 4) As the pump is placed in the aorta, it does not cross the aortic valve and, therefore, will not come into conflict with a diseased aortic valve or aortic valve prosthesis. Hence, the pump can be used even in cases of aortic insufficiency, aortic stenosis, or aortic valve prosthesis.
A common axial flow pump has impellers and a housing with a defined diameter, the diameter of the impeller being only slightly less than the internal diameter of the housing. In contrast to this, the present pump has a variable and larger gap between the propeller (impeller) and the tube (i.e., aortic wall) for two reasons. First, a fixed gap is created between the tip of the propeller and the expanded cage, and second, a variable gap is found between the cage and the aorta. The total gap will, therefore, lead to a backward leakage of fluid along the wall of the tube, which tends to reduce the hydraulic effect of the pump.
We have previously reported results from a bench test in which different pump catheters without a cage and with different propellers were tested. 9 Both the form of the cage and the profile of each filament will tend to reduce the performance of the propeller. Despite this reduction, we have observed significant pressure gradients to occur in healthy animals, as well as in animals with heart failure (to be published) (Figure 2).
The aim of this study was to test the hydrodynamic properties of the latest version of the pump (February 1998) with respect to 1) revolutions per minute, propeller sizes (diameters), and the tube size on pressure generation; 2) calculate the degrading effect of the cage; and 3) describe the flow-pressure relationship.
Materials and Methods
Design of the Pump
The pump head with the foldable propeller and cage is opened and closed by means of an umbrella-like mechanism at the drive unit, with the motion transmitted by means of a flexible catheter. The catheter consists of two elements, one outer and one inner sleeve. A longitudinal displacement between the two parts transfers the movement to the opening mechanism for the propeller and cage. The opposite movement closes the propeller and the cage and enables the removal of the catheter pump from the insertion site. The rotation of the propeller is transmitted by means of a wire inside the inner lumen of the catheter connected to a DC motor at the proximal end in the drive unit. The pump head has two bearings, one placed proximally at the hinges of the filaments. To reduce friction and heat generation of this bearing, it is lubricated with a mixture of water and glucose that is transported from the drive unit to the pump head by means of narrow channels inside the catheter. The distal bearing is immersed in blood and has no active lubrication. The propeller, made of stainless steel, consists of two separate hinged components, each forming a wing. The diameter of the opened propeller in the present pump is 15 mm and the hydraulic area of each wing is 3 × 4 mm on the outer part with an outer angle of 45 degrees. The 3.5 mm of the inner part is left for the hinge system to open and close the propeller. The deployed cage has an outer diameter of 21 mm. A console controls the purge fluid and revolutions per minute of the pump. The revolutions per minute can be finely adjusted between 0 to a maximum of 15,000 revolutions per minute (rpm).
The pumps were tested with a mixture of 47% glycerol and water as test fluid. This mixture has a viscosity of 3.4 × 10−3Ns/m2, which is close to the viscosity of blood. Blood has a viscosity of approximately 3.5 × 10−3Ns/m2, with an erythrocyte volume fraction of 40%. 10 The density of the fluid was 1,113 g/cm3, and its temperature was 22°C during the tests.
Two different bench models were used. The first model consisted of a vessel with fluid and exchangeable tilted acrylic tubes, with one end of the tube in the fluid. This model allowed for testing the pressure generating properties of the pumps at zero flow with various revolutions per minute and propeller sizes, by measuring the height of the column of fluid inside the tube. This system is called the pressure model. The second model was a nonpulsatile mock-loop with a flowmeter and an adjustable resistance to obtain the flow-pressure relationship (called the loop) (Figure 3, A and B). The pump was placed in a part of the loop with exchangeable tubes, so that the influence of different diameters of the tube could be tested. Because the normal size of the human ascending aorta ranges from 20 to 40 mm in diameter (15–30 mm is the usual range found in the descending aorta), the tubes tested were 22–30 mm with 2 mm increments.
Two different pumps were tested. In the pressure model, one pump with different propellers but without a cage was tested to obtain the relationship between the diameter of the propeller and generated pressure. The propellers were made of bent metal plates with a width of 3 mm and a thickness of 0.4 mm, corresponding to the width and thickness of the caged version of the pump. The angle of the propellers was fixed to 45 degrees. The propellers used were 14–20 mm with 1 mm increments. In the pressure model a semiquantitative test was also performed to examine the degrading effect of the cage comparing the pressure generation of two different pumps, one with and one without a cage. The caged pump was the latest prototype of the complete pump and had been used in the animal trials.
In the loop, flow-pressure tests were done with the complete pump to examine 1) the effect of different revolutions per minute on the pressure generation at zero flow, 2) the pressure degradation by increasing the size of the tube at zero flow, and 3) the flow-pressure relationship in the different tubes at constant revolutions per minute. Curve fitting and regression analyses were performed with commercial software (Lotus Freelance Graphics 97). Results are given as mean ± 1 SD.
The comparison between the two pumps in the pressure model showed a pressure reduction of 15–20% for the complete pump because of the cage (Figure 4). It was also found that there was a diameter dependent rise in the generated pressure with a larger propeller at a constant revolutions per minute according to the equation:
MATH where P0 is the pressure at zero flow, c1 is a constant, Dp is the diameter of the propeller, and b1 is the power to Dp. The exponent (b1) was found to be 3.92 (± 0.18 n = 22) (Figure 5).
In the loop, the pump showed a revolutions per minute dependent rise in pressure at zero flow, according to the equation (Figure 4):MATH where P0 is the generated pressure, and c2 is a constant. The exponent (b2) was found to be 1.99 (± 0.08, n = 5).
At a constant revolutions per minute and zero flow, there was a reduction in the pressure gradient with increasing size of the tube (Figure 6):
MATH where P0 is the generated pressure, and c3 × Dt the diameter of the tube. In the range 9,000 to 15,000 rpm, the exponent (b3) was found to be −2.7 (± 0.08 n = 7).
At constant revolutions per minute, the flow-pressure relationship showed a nearly linear pressure fall, with an increasing flow in the two smallest tubes. However, the pressure reduction decreased with increasing size of the tube, and for the 30-mm tube, practically no pressure fall was seen to a flow of 10 L (Figure 7).
We have studied the hydrodynamic effects of a new intra-aortic axial flow pump in a bench model. The tests were carried out with a mixture of water and glycerol with a viscosity similar to that of blood. Earlier pump prototypes with propellers of different sizes have been tested in animals in both the ascending and descending parts of the aorta. 9 Significant pressure gradients of up to 48 mm Hg could be achieved in vivo, the magnitude of the pressure gradient being dependent upon the rotational speed and size of the propeller. The design of the pump is based on a foldable construction and 4.8 mm was decided on for the outer diameter of the closed pump head. This design also reduces the size of the propeller to a maximal width of 3 mm to fit into the construction. Consequently, there are limited possibilities for optimization of the proportions of the propeller dimensions with respect to hydrodynamic properties. Increasing the width of the propellers would entail increased thickness of the folded pump head, whereas increasing the length of the propeller wings would not.
The pump requires a cage, therefore, it was important to examine the degrading effect of the cage as one of the test series was done with a cageless pump. The ultimate goal is to construct a cage with a negligible degrading effect. In this study, the present cage design was found to reduce the propeller effect by 15–20% with respect to pressure, compared with 40% with an earlier cage design. 9
In the pressure model at zero flow, we identified three major factors that influence the generated pressure. The most important is the diameter of the propeller, because this factor was found to influence the pressure to the 3.92nd power at constant revolutions per minute. A change in dimension from 15 to 16 mm would increase the generated pressure by 30% at constant revolutions per minute. The second factor was the rotational speed, which was found to influence the generated pressure to the second power of the revolutions per minute. 11 These two factors, in addition to the cage, can all be technically modified. The third important factor influencing the generated pressure was the diameter of the tube in vitro or the diameter of the aorta in vivo. With increasing size of the tube, the generated pressure was reduced to the 2.7th power of the diameter.
Although the diameter of the aorta has a great impact on performance of the pump, it is impossible to modify and makes it difficult to determine the definite performance characteristics of the pump in all patients. This is unlike other axial flow pumps where the outer housing is kept constant and the effect is predictable. On the other hand, the mathematical relationship found in the present study makes it possible to calculate the pressure effect for different diameters of tubes or aorta within certain limits (Figure 6). One limitation of such calculations is the compliance of the aorta, causing changes in its diameter of 1–2 mm from diastole to systole. 12 A second is a small pressure drop with increasing blood flow.
In the loop model, we observed a nearly linear relationship between pressure and flow with the smaller tubes. However, the pressure drop was small, indicating that the pump is able to move great amounts of fluid. Interestingly, no pressure drop occurred with a flow of up to 10 L/min when using bigger tubes, probably because the back-flow along the tube wall at zero flow was disturbed with the increasing flow.
For a pump placed in the aorta, it is important that it is able to receive large volumes with a minimal pressure drop, otherwise it could act as an obstruction at low revolutions per minute. Systole occupies one third of the total heart cycle in a healthy person. With a cardiac output of 5 L/min, this value would correspond to a continuous flow rate of approximately 15 L/min, dropping to 5–10 L in heart failure. 13
In a recent study with a series of calves in which the pump was positioned in the descending aorta, there was a significant pressure drop proximal to the propeller and a more pronounced increase in the distal pressure. In all animals, a pressure gradient across the pump was measurable at a rotational speed above 6,000 rpm (right femoral pressure minus right carotid pressure). 14 This demonstrates that the pump does not obstruct the flow in vivo in the aorta, even at a low rotational speed.
What is the minimal pressure gradient that could be of clinical value? Experience with drug induced afterload reduction in patients with left ventricular failure tells us that even a small reduction in the arterial pressure has a beneficial effect on cardiac performance. Increased cardiac output is commonly found with a pressure reduction of 5 mm Hg. 15–17 However, a pharmacologically induced afterload reduction is often limited by hypotension causing peripheral hypoperfusion. The consequences of hypotension are multiorgan failure, including renal dysfunction and bowel ischemia. 18–20 The catheter pump might achieve a favorable proximal pressure reduction, and at the same time maintain or increase peripheral perfusion. A pressure gradient of 10 mm or more, therefore, is probably clinically relevant in addition to pharmacologic treatment.
With the intra aortic balloon pump, which is also placed in the descending aorta, ventricular systolic pressure drops between 2 and 10 mm Hg have been reported when tests are done in a state of a regular cardiac rhythm, either sinus or paced. 21,22 In atrial fibrillation, however, the systolic unloading will vary and be reduced because of the variable filling of the left ventricle and the filling of the IABP itself. 23 In contrast to the IABP, the present pump is not triggered by the heart and works independent of the cardiac cycle and, therefore, will probably not be influenced by cardiac arrhythmia to the same extent. Distal to the present pump, we have always observed an increased pressure, whereas some reports show that the IABP may sometimes reduce the pressure in the lower part of the body as well as reduce renal flow. 24,25
Long-term tests of hemolysis (24 hour) have been performed in calves with earlier prototypes of the pump as well as with the present one. They have shown that the pump caused hemolysis up to 1,000 mg of hemoglobin per liter of plasma at rotational speeds of up to 14,000 rpm over a period of 24 hours. However, ongoing tests with a new propeller design indicate that the hemolysis rate can be markedly reduced, in addition to improved efficacy of the propeller. Therefore, we expect that the number of revolutions chosen in this study can be used clinically with an acceptable degree of hemolysis.
During the development of the pump, different bearings at the tip have been tested because of problems with heat generation and fibrin formation. After a change of bearings, the problem with both heat generation and thrombus formation seems to be solved. When the final design has been decided upon, more extensive testing of biocompatibility will be performed.
In conclusion, the initial experience with this new, propeller driven intra-aortic axial flow pump shows promise. Our results demonstrate that the catheter pump has hydrodynamic properties that might allow for its use as a circulatory support device in a clinical setting.
1. Kasper EK, Agema WR, Hutchins GM, et al: The causes of dilated cardiomyopathy: A clinical review of 673 consecutive patients. J Am Coll Cardiol 23:586, 1994.
2. Dzavic V, Burton JR, Kee C, et al: Changing practice patterns in the management of acute myocardial infarction complicated by cardiogenic shock: Elderly compared with younger patients. Can J Cardiol 14:923–930, 1998.
3. Pae WE Jr., Miller CA, Mattews Y, et al: Ventricular assist devices for postcardiogenic shock: A combined registry experience. J Thorac Cardiovasc Surg 104:451–453, 1992.
4. Hill JG, Bruhn PS, Cohen SE, et al: Emergent application of cardiopulmonary support: A multi-institutional experience. Ann Thorac Surg 54:699–704, 1992.
5. DeRose JJ Jr., Umana JP, Argenziano M, et al: Improved results for postcardiotomy cardiogenic shock with the use of implantable left ventricular assist devices. Ann Thorac Surg 64:1757–1763, 1997.
6. Christenson JT, Buswell L, Velebit V, et al: The intraaortic balloon pump for postcardiotomy heart failure. Experience with 169 intraaortic balloon pumps. Thorac Cardiovasc Surg 43:129–133, 1995.
7. Peterzen B, Lonn U, Babic A, et al: Management of patients with Hemopump support after coronary artery bypass grafting. Ann Thorac Surg 62:495–500, 1996.
8. Orime Y, Shindo S, Shiono M, et al: Experiences of postcardiotomy assist: pneumatic ventricular assist device or veno-arterial bypass with percutaneous cardiopulmonary support. Artif Organs 20:721–723, 1996.
9. Reitan Ö, Öhlin H, Peterzén B, et al: Initial tests with a new cardiac assist device. ASAIO J 45:317–321, 1999.
10. Urban L, Wulff J, Keck KY, et al: Flow characteristics of the Hemopump: An experimental in vitro study. Ann Thorac Surg 63:162–166, 1997.
11. White FM: Mixed and axial-flow pumps, in White FM (ed), Fluid Mechanics. New York, Schaum Division McGraw-Hill, Inc., 1994, p. 664.
12. Pearson AC, Guo R, Orsinelli DA, et al: Transesophageal echocardiographic assessment of the effects of age, gender, and hypertension on thoracic wall size, thickness, and stiffness. Am Heart J 128:344–351, 1994.
13. Ganong WF: Systolen- und Diastolen-Dauer, in Medizinische Physiologie. Berlin, Heidelberg, New York, Springer Verlag, 1972, p. 519.
14. Reitan Ö, Steen S, Öhlin H: Tests with a percutaneous cardiac assist device in a left ventricular failure model. ASAIO J 45:170A, 1999.
15. Guiha NH, Cohn JN, Miculic E, et al: Treatment of refractory heart failure with infusion nitroprusside. N Engl J Med 291:587–592, 1974.
16. Yin FCP, Guzman PA, Brin KP, et al: Effects of nitroprusside on hydraulic vascular loads on the right and left ventricle of patients with heart failure. Circulation 67:1330–1339, 1983.
17. Parmley WW, Chatterjee K, Charuzi Y, et al: Hemodynamic effects of noninvasive systolic unloading (nitroprusside) and diastolic augmentation (external counterpulsation) in patients with acute myocardial infarction. Am J Cardiol 33:819–825, 1974.
18. Suen WS, Mok CK, Chiu SW, et al: Risk factors for development of acute renal failure (ARF) requiring dialysis in patients undergoing cardiac surgery. Angiology 49:789–800, 1998.
19. Zanardo G, Michielon P, Paccagnella A, et al: Acute renal failure in the patient undergoing cardiac operation: Prevalence, mortality rate, and main risk factors. Thorac Cardiovasc Surg 107:1489–1495, 1994.
20. Christenson JT, Schmuziger M, Maurice J, et al: Postoperative visceral hypotension the common cause for gastrointestinal complications after cardiac surgery. Thorac Cardiovasc Surg 42:152–157, 1994.
21. Smalling RW, Cassidy DB, Barrett R, Lachterman B, Felli P, Amirian J: Improved regional myocardial blood flow, left ventricular unloading, and infarct salvage using an axial-flow, transvalvular left ventricular assist device: A comparison with intra-aortic balloon counterpulsation and reperfusion alone in a canine infarction model. Circulation 85:1152–1159, 1992.
22. Cheung AT, Savino JS, Weiss SJ: Beat to beat augmentation of left ventricular function by intraaortic counterpulsation. Anesthesiology 84:545–554, 1996.
23. Sakamato T, Arai H, Maruyama T, Suzuki A: New algorithm of intra-aortic balloon pumping in patients with atrial fibrillation. ASAIO J 41:79–83, 1995.
24. Baldwin RT, Radovancevic B, Conger JL, et al: Peripheral organ perfusion augmentation during left ventricular failure: A controlled bovine comparison between the intraaortic balloon pump and the Hemopump. Tex Heart Inst J 20:275–280, 1993.
Copyright © 2000 by the American Society for Artificial Internal Organs
25. Swartz MT, Sakamoto T, Arai H, et al: Effects of intra-aortic balloon pumping on renal artery blood flow. Ann Thorac Surg 53:604–610, 1992.