Ann Thorac Surg 1997;63:162-166
© 1997 The Society of Thoracic Surgeons
Original Articles: Cardiovascular
Flow Characteristics of the Hemopump: An Experimental In Vitro Study
Urban Lönn, MD,
John Wulff, MD,
Karl-Yngve Keck, MSEE,
Bengt Wranne, MD, PhD,
Per Ask, MSEE, PhD,
Bengt Peterzén, MD,
Henrik Casimir-Ahn, MD, PhD
Departments of Cardiothoracic Anesthesiology, Cardiothoracic Surgery, Physiology, and Biomedical Engineering, Linköping Heart Center, University of Linköping, Linköping, Sweden
Accepted for publication July 24, 1996.
 |
Abstract
|
|---|
Background. The Hemopump (DLP/Medtronic) has been in clinical use for about 7 years. There is still no adequate way of determining actual output from the three available pump systems in the clinical situation. If the pump is completely stopped during weaning from the device, there is a possibility of back-leakage through the pump, endangering the patient from regurgitation into the left ventricle. It can also make it more difficult to judge the recovery of heart function because of a volume load of the left ventricle. The aim of this study was to evaluate in a standardized, experimental in vitro model the output from three different-sized Hemopump catheters at various pressure levels and to quantify the back-flow through the pumps.
Methods. The Hemopump models were tested in an in vitro study regarding total outflow at various speeds at three pressure levels. The back-flow through the pumps was also measured with the pumps at a complete stop.
Results. The outflow from the Hemopumps ranged from 0.4 to 4.5 L/min, depending on which pump and speed were used. Variations in total output, depending on speed and various pressure settings, could be up to 0.4 L/min. Back-flow through the pump into the left ventricle may be as great as 1.6 L/min.
Conclusions. The flow outputs from the different Hemopump models were reproducible over time and were closely related to the resistance of the model. The Hemopump, if not running, can induce substantial regurgitation through the pump into the left ventricle.
|
Introduction
|
|---|
The Hemopump (HP) (DLP/Medtronic Inc, Grand Rapids, MI) is a small axial turbo blood flow pump intended to be used as a temporary left ventricular assist device in patients who have potentially reversible left ventricular failure. The pumps are available in three different dimensions (Fig 1
): (1) A 14F pump (HP-14) designed to be placed percutaneously through the femoral artery; (2) A 21F pump (HP-21), placed in position through a small cutdown into the femoral artery, either through a graft or using pursestring sutures; and (3) A 21F pump (HP-31) that is a further development of the HP-21 and is designed to be used during open heart operations. The pump is put in place through a graft sutured to the ascending aorta.
The function of the pump is to unload the left ventricle by aspirating blood and pumping it into the aorta. This is achieved by a catheter with its distal end placed in the left ventricle through the aorta. The pump itself is positioned at the proximal end of the catheter (Fig 2). The drive console allows adjustment of the speed of the pump in seven increments, where "1" is the minimum speed and "7" is the maximum speed (Fig 3). The console controls give no measure of the actual blood flow delivered by the pump.
View larger version (97K):
[in this window]
[in a new window]
|
Fig 2. . The Hemopump HP-21 femoral catheter in position within the heart, and its function. (Courtesy of Medtronic Inc, Hemodynamics Division, Grand Rapids, MI.)
|
|
View larger version (122K):
[in this window]
[in a new window]
|
Fig 3. . The Hemopump device with drive unit, with the HP-31 connected, and with the electromagnetic motor. A slow, continuous infusion of glucose is used for cooling and lubricating the pump. (Reprinted by permission of The Society of Thoracic Surgeons [Ann Thorac Surg 1995;59:S36–45].)
|
|
We have used these devices since 1992 and have been impressed by the system's capability of unloading the failing left ventricle, giving the heart what we consider to be an excellent chance of recovery without using excessive levels of inotropic support [1–4]. Other situations in which the HP has been used include support for patients undergoing high-risk coronary angioplasty [5], as treatment of allograft failure [6], and as circulatory support instead of the heart-lung machine during coronary artery bypass grafting [7, 8].
In our experience, cardiac output measurement using the thermodilution technique has not been reliable in assessing pump function, because it cannot discriminate between the two pumping systems: that of the assist device, and the heart's own contribution. One of the aims of this study was to quantify, in a standardized in vitro model, the flow at different pressure levels.
When weaning patients from the HP, we have observed various degrees of regurgitation through the pump if it is stopped abruptly. It has been difficult to estimate the hemodynamic effects of this regurgitation. Therefore, the possibility of back-flow through the three pumps was also studied in the same model.
|
Material and Methods
|
|---|
Operating Principle of the Hemopump
The HP has been in clinical use since 1990, when Frazier and associates [2] described their first clinical experience in the treatment of 7 patients with cardiogenic shock. The HP gives a nonpulsatile blood flow independent of the heart's own rhythm. The system is designed to unload the left ventricle by aspirating blood from the ventricle and pumping it into the aorta. The construction consists of a small axial-flow pump placed in a short tube that is open at both ends. The pump is driven by a metallic wire, which is covered by a plastic sheet. The wire is rotated by a paracorporeal electromagnetic motor, controlled by a small drive console. The speed of the pump can be adjusted from 17,000 to 48,000 revolutions/min depending on which catheter is used. The drive wire is lubricated with dextrose, which is purged into the plastic sheet. Placement of the pump is achieved by gaining access to the aorta through the femoral artery or through the ascending aorta. The distal end of the catheter is then placed through the valve into the left ventricle, enabling blood to be aspirated from the ventricle and delivered into the aorta (see Fig 2
).
In Vitro Model
The flow characteristics of each pump were evaluated in an in vitro model (Fig 4). This consisted of a container filled with the fluid to be pumped (A). A tube was connected to this container through which the HP was introduced. The tube was sealed around the HP so that no back-flow could occur besides the HP catheter down to container A.
View larger version (22K):
[in this window]
[in a new window]
|
Fig 4. . Experimental model. Liquid is pumped by the Hemopump (HP) through a tube from container A up to container B. A seal between the HP cannulas and the tube prevents back-flow besides the HP. The height between the two containers can be adjusted to create different pressures toward the HP outlet. Stopping the pump will create back-flow through the HP cannulas from container B down to container A.
|
|
Distal to the HP, this system was connected to a second container (B), the height of which was adjustable. The pressure against which the HP had to work was regulated by changing the height of container B relative to A. The fluid to be pumped consisted of glycerine 45.7% by weight in water, the viscosity of which was measured to be 3.4 x 10-3 Ns/m2. Blood with an erythrocyte volume fraction of 40% and at normal temperature has a viscosity of 3.5 x 10-3 Ns/m2. The density of the fluid was 1,113 kg/m3. The ambient temperature and that of the fluid were 21°C.
Three different levels of pressure were used: low pressure, intermediate pressure, and high pressure (Table 1). The pressure was measured with a portable pressure monitor (M 1275 A; Hewlett-Packard) connected to a pressure transducer set (TriPlus; Kirchseeon, Germany). Flow was measured using a Transonic ultrasound transit-time flowmeter (HT 109; Transonic System Inc, Ithaca, NY).
Three separate flow and pressure readings were done in each experiment. Back-flow was measured by stopping the HP and letting the fluid pass backwards through the pump housing from container B down to container A.
|
Results
|
|---|
Results are given as mean ± standard deviation.
Flow Characteristics of the Hemopump-14
In the low pressure setting (29.4 ± 0.5 mm Hg; n = 21), the HP-14 gave at minimum speed (level 1) a flow of 0.85 ± 0.02 L/min, and at maximum speed (level 7) 1.96 ± 0.01 L/min (Fig 5). Corresponding values for the intermediate pressure setting (43.7 ± 0.5 mm Hg; n = 21) were 0.81 ± 0.01 L/min and 1.91 ± 0.03 L/min. In the high pressure mode (63.4 ± 0.8 mm Hg; n = 21), the HP gave a flow at level 1 of 0.45 ± 0.01 L/min and at level 7 of 1.63 ± 0.02 L/min.
View larger version (62K):
[in this window]
[in a new window]
|
Fig 5. . Flow profiles of the different Hemopumps (HP) at the various speeds and pressure settings. (Inter. = intermediate; P = pressure.)
|
|
The back-flow (Fig 6) through the pump was linearly related to the pressure (P) level and was given by the following equation: Q HP-14 = 0.30 + (0.01P); r = 0.99.
View larger version (25K):
[in this window]
[in a new window]
|
Fig 6. . Back-flow through the Hemopumps (HP) with the pump in the "off" mode at three different pressure levels.
|
|
Flow Characteristics of the Hemopump-21
In the low pressure setting (29.1 ± 0.8 mm Hg; n = 21), the HP-21 gave at minimum speed (level 1) a flow of 1.76 ± 0.04 L/min, and at maximum speed (level 7) 3.06 ± 0.02 L/min (see Fig 5). Corresponding values for the intermediate pressure setting (43.6 ± 0.5 mm Hg; n = 21) were 1.51 ± 0.02 L/min and 2.91 ± 0.01 L/min. In the high pressure mode (63.3 ± 0.46 mm Hg; n = 21), the HP gave a flow at level 1 of 1.08 ± 0.01 L/min and at level 7 of 2.74 ± 0.02 L/min.
The back-flow (see Fig 6) through the pump was linearly related to the pressure (P) level and was given by the following equation: Q HP-21 = 0.37 + (0.012P); r = 0.99.
Flow Characteristics of the Hemopump-31
In the low pressure setting (32.0 ± 1.6 mm Hg; n = 21), the HP-31 gave at minimum speed (level 1) a flow of 2.55 ± 0.04 L/min, and at maximum speed (level 7) 4.44 ± 0.01 L/min (see Fig 5). Corresponding values for the intermediate pressure setting (45.1 ± 1.1 mm Hg; n = 21) were 2.16 ± 0.01 L/min and 4.38 ± 0.01 L/min; values in the high pressure setting (64.8 ± 1.0 mm Hg; n = 21) were 1.28 ± 0.01 L/min and 4.08 ± 0.01 L/min, respectively.
Back-flow (see Fig 6) through the pump again was linearly related to the pressure (P) level and was given by the following equation: Q HP-31 = 0.53 + (0.016P); r = 0.99.
|
Comment
|
|---|
Maximum outflows from the different HP catheters have been claimed to be 2 L/min, 3.5 L/min, and 5.5 L/min for the HP-14, HP-21, and HP-31 systems, respectively. We have observed in the clinical situation using the HP-21 and HP-31 pumps in patients with postcardiotomy heart failure that the patient's arterial blood pressure recording initially demonstrates a completely flat line without pulsation [1]. Because the HP generates a nonpulsatile blood flow, this suggests that the aortic valves are not opening and that all blood flow will pass through the HP. We have been able to verify this using echocardiography. The use of the thermodilution technique in an attempt to quantify and verify the output from the HP in the clinical situation has not been reliable; measured values have varied widely from those expected. The reason for this is unknown. There is no way of estimating flow output from the HP itself in vivo. It is reasonable to assume that mean arterial blood pressure will influence the pump's maximum output. We therefore believed it was necessary to develop a standardized in vitro model to examine these variables. We know from the clinical situation that a low peripheral resistance is desirable, not only for the purpose of good organ perfusion, but also to provide optimal conditions for the pump to unload the heart and give maximum output. We have previously reported that mixed venous saturation together with urinary output and mean arterial pressure are the most important factors to follow in patients undergoing HP treatment [1].
Our results indicate that abrupt stopping of the pump may lead not only to acute left ventricular failure, but also to the appearance of regurgitation from back-flow through the HP. This together with the risk for thromboembolism formation gives reason to recommend not stopping the pump completely under any circumstance. When evaluating the capability of a patient to maintain adequate circulation without the help of this device, we now routinely decrease the degree of assistance in stages over a period of at least 6 hours, until we reach speed level "1." We then keep the pump at the level for a further 2 hours before the pump is removed, in case the patient should require additional support. The HP will contribute significantly to the circulation at the setting if the HP-31 is used. It gives 2.16 L/min at a pressure level of 45 mm Hg, and the contribution from the pump is 1.3 L/min at a pressure of 65 mm Hg. It is important to be aware of this when evaluating recovery of the heart. It would be of great value to have some way of setting the pump speed to a level corresponding to the mean arterial pressure, so that we would know that the flow through the HP would be close to zero. It would also be a big advantage to be able to monitor blood pressure at the tip of the pump catheter. These two suggestions should help to evaluate the correct position in the left ventricle when introducing the device and, during a weaning situation, to be able to follow the filling pressure in the left ventricle.
In our early experience with the HP, we had 1 patient who showed signs of recovery and was weaned from the device in stages. Finally the pump was stopped, and ventricular recovery was judged from ultrasonic echocardiography. Based upon these findings, the pump was removed. Soon after that, the patient started to deteriorate because of heart failure. In retrospect, what we saw was a left ventricle with lively and rapid contractions caused by volume overload due to regurgitation through the HP, and not true ventricular recovery.
The flow characteristics of the HP, as described in this report, may help explain the findings reported by Dubois-Randé [9]. They evaluated the HP-14 in patients undergoing high-risk coronary angioplasty. They stated that the HP-14 gives a flow rate of 2.5 to 3.0 L/min and, based upon their observations, concluded that the pump could unload the left ventricle. In their 10 high-risk patients, the pulmonary capillary wedge pressure was 29 ± 7 mm Hg with the catheter positioned in the left ventricle and with the pump off. When they then put the pump on at maximum speed, the pulmonary capillary wedge pressure decreased to 22 ± 8 mm Hg during the coronary angioplasty procedure. The mean arterial pressure was 72 mm Hg with the pump off. In light of our observations, the hemodynamic picture presented by Dubois-Randé and colleagues [9] could well be explained by regurgitation into the left ventricle through the pump. At a mean arterial pressure of 72 mm Hg, the back-flow through the HP-14 in the "off" mode could be as much as 1.0 L/min. The elevated pulmonary capillary wedge pressure, as we see it, should be explained by regurgitation, and the difference in wedge pressures should not be taken as a sign of ventricular unloading in the situation presented. The blood flow given by the HP-14 at pressures greater than 70 mm Hg is well below 2.0 L/min. This is also consistent with the technical limitations described by Ruel [10] for miniaturized axial flow pumps.
In conclusion, awareness of the results described in this report will help clinicians to treat patients with the HP system. We have successfully treated patients with potentially reversible cardiac failure using this device, as described previously [1]. The system, if used correctly, shows great potential and should be considered as a valuable tool in the management of patients with a failing left ventricle.
|
Acknowledgments
|
|---|
This study was supported by the Swedish Medical Research Council (grant 9481), The Swedish Heart and Lung Foundation, and the Swedish Research Council for Engineering Science. We thank Per Sveider for building the experimental model.
|
Footnotes
|
|---|
Address reprint requests to Dr Lönn, Department of Thoracic Surgery, Linköping Heart Center, University of Linköping, Linköping, Sweden.
|
References
|
|---|
- Lönn U, Peterzén B, Granfelt H, Babic A, Casimir-Ahn H. Hemopump treatment in patients with postcardiotomy heart failure. Ann Thorac Surg 1995;60:1067–71.[Abstract/Free Full Text]
- Frazier OH, Wampler RK, Dunkan JM, Dear WE, Macris MP. First human use of the Hemopump, a catheter-mounted ventricular assist device. Ann Thorac Surg 1990;49:299–304.[Abstract/Free Full Text]
- Burnett CM, Vega JD, Radovancevic B, Longquist JL, Birovljev S. Improved survival after Hemopump insertion in patients experiencing postcardiotomy cardiogenic shock during cardiopulmonary bypass. ASAIO Trans 1990;36:M626–9.[Medline]
- Wampler RK, Frazier OH, Lansing AM, et al. Treatment of cardiogenic shock with the Hemopump left ventricular assist device. Ann Thorac Surg 1991;52:506–13.
- Loisance D, Deleuze P, Dubois-Randé JL, Okude J, Shiiya N. Hemopump ventricular support for patients undergoing high risk coronary angioplasty. ASAIO Trans 1990;36:623–6.
- Frazier OH, Macris MP, Wampler RK, Dunkan JM, Sweeney MS. Treatment of cardiac allograft failure by use of an intraaortic axial flow pump. J Heart Transplant 1990;9:408–14.[Medline]
- Lönn U, Peterzén B, Granfelt H, Casimir-Ahn H. Coronary artery operation supported by the Hemopump: an experimental study on pig. Ann Thorac Surg 1994;58:516–8.[Abstract/Free Full Text]
- Lönn U, Peterzén B, Granfelt H, Casimir-Ahn H. Coronary artery operation with support of the Hemopump cardiac assist system. Ann Thorac Surg 1994;58:519–23.[Abstract/Free Full Text]
- Dubois-Randé JL, Deleuze P, Dupouy P, Geschwind H, Loisance D. Assessment of a percutaneous Hemopump in high risk coronary angioplasty patients. ASAIO Trans 1994;40:486–8.
- Reul H. Technical requirements and limitations of miniaturized axial flow pumps for circulatory support. Cardiology 1994;84:187–93.[Medline]
This article has been cited by other articles:
|
|
|
|
 
M. D. Rodefeld, J. H. Boyd, C. D. Myers, B. J. LaLone, A. J. Bezruczko, A. W. Potter, and J. W. Brown
Cavopulmonary assist: circulatory support for the univentricular fontan circulation
Ann. Thorac. Surg.,
December 1, 2003;
76(6):
1911 - 1916.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. Schmidt, J. Siefker, S. Spiliopoulos, and O. Dapunt
New experience with the paracardial right ventricular axial flow micropump impella elect 600
Eur J Cardiothorac Surg,
August 1, 2003;
24(2):
307 - 308.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
