Secondary Logo

Journal Logo


Left Ventricular Heart Failure Model for Testing Cardiac Assist Devices

Reitan, Öyvind,*; Steen, Stig,† and; Öhlin, Hans*

Author Information
  • Free


Mechanical support of the failing heart has been a widely used treatment for many years, and there is a growing number of new devices. 1–3 The development of a new device usually requires animal testing before human use to ensure efficacy and safety.

Most cardiac assist systems support either the left or the right ventricle. They are intended to treat serious heart failure when medical treatment fails, and, consequently, the best conditions for testing new devices are in animals with heart failure. The normally working heart is not optimal for the examination of a support system with respect to hemodynamic response; nevertheless, it offers the biologic environment needed for testing durability and safety such as thromboembolic complications and damage to blood elements. 4,5

We are currently developing a circulatory support system for the left ventricle and have used animals with and without heart failure to evaluate the hemodynamic effect. 6,7 In animals with normal heart function, it has been difficult to judge the hemodynamic contribution of the device.

Many different animal models of heart failure have been published. Induced heart infarction by ligation of one or more coronary arteries will produce heart failure, but the degree is somewhat unpredictable. 8–10 Injection of microspheres into the coronary arteries is another method and produces microinfarctions over larger areas of the left or right ventricle. 11,12 Rapid ventricular pacing can also be used to bring on heart failure. Pacemaker induced tachycardia with a pulse rate of more than 120–130 beats per minute disturbs the atrial-ventricular synchrony and the filling of the left ventricle and leads to biventricular failure, especially during beta blockade. 13

Our own experience with induced myocardial infarction in pigs has revealed problems with varying degrees of heart failure, sudden electric and hemodynamic instability, and, sometimes, death of animals during the tests. A failure model based on the negative inotropic effect of some drugs normally leads to biventricular failure and is not found ideal for testing a left ventricular support device. 14–16 One of the key features of left ventricular heart failure is the presence of high filling pressures of the left ventricle 17 with the consequence that the mean left atrial pressure (LAP) will rise because of backward congestion. Therefore, our aim was to create an adjustable left ventricular failure model in which LAP could be varied for acute hemodynamic testing of our other support systems.

Materials and Methods

Anesthesia and Operative Management

Seven domestic calves with a mean weight of 84 kg (range, 70–95 kg) were used in the experiments, which were approved by the local ethical committee. The animals received humane care in compliance with the “Guide for the Care and Use of Laboratory Animals” (NIH publication no. 85–23, revised 1985).

Anesthesia was induced with 1,000 mg of ketamine (50 mg/ml Ketalar; Parke Davis, Morris Plains, NJ) given intramuscularly. Thiopental sodium 25 mg/ml (Pentothal; Abbott Laboratories, North Chicago, IL) at a dose of 10 mg/kg and atropine (Kabi Pharmacia, Uppsala, Sweden) at a dose of 0.5 mg were given intravenously before tracheostomy (no. 8.5 tube). Anesthesia and muscle relaxation were maintained with a continuous infusion of fentanyl at a dose of 0.1 mg/hour (50 μg/ml; Leptanal, Janssen Cilag) and pancuronium bromide (2 mg/ml; Pavulon, Organon Teknika, Boxtel, The Netherlands), given at a rate of 0.1 mg/kg per hour. The animals were ventilated with a Siemens Servoventilator 900 (Siemens-Elema AB, Stockholm, Sweden). A volume controlled, pressure regulated ventilation of 10 L/min (20 breaths per min; positive end-expiratory pressure 8 cm H2O; inspired oxygen fraction, 0.5) with 1–2% isoflurane (Abbott Scandinavia, Stockholm, Sweden) was used.

A Foley catheter was inserted into the urinary bladder by means of a suprapubic cystostomy. A right neck incision was made and an arterial line introduced into the right carotid artery and connected to a pressure transducer (Ohmeda 1DT 1Rose, Murray Hill, NJ) to monitor the proximal arterial pressure. A right femoral incision was made to expose the femoral artery, and an arterial line was established and connected to a pressure transducer to monitor the distal arterial pressure. A Doppler flow transducer was connected to the right femoral artery (Transonic T201, Ithaca, NY) to measure the femoral artery flow. The left carotid artery was exposed by means of a left neck incision, and a flow transducer was connected to the artery to measure the carotid flow (Transonic T201). A left sided lumbar incision was made, the left renal vein was exposed, and a Doppler probe (Transonic T201) was connected to measure renal blood flow. Recordings of femoral and renal blood flow and femoral pressure were followed when the animal was on the support device.

A sternotomy was performed, the heart was exposed, and the pulmonary trunk was dissected free; a flow transducer (Transonic) was connected to the pulmonary artery to enable continuous measurement of cardiac output of the healthy animal. The proximal left anterior descending artery (LAD) was dissected and equipped with a flow transducer (Transonic).

The LAP was measured with a catheter through the left atrial appendage, and the left ventricular pressure (LVP) was measured by means of a line through the ventricular wall. The right heart was similarly equipped, with the right atrial pressure measured by means of a catheter through the right atrial appendage and the right ventricular pressure measured by means of a catheter through the anterior ventricular wall. The pressure in the pulmonary artery was measured with a line into the vessel.

Failure Model

The present heart failure model is based on the negative inotropic effect of pharmacologic agents with well known cardiodepressive effects. We used a combination of a beta-blocker and a calcium antagonist. Metoprolol, 10–25 mg (mean, 17.3 ± 4.8 mg), was administered intravenously as a bolus dose to block the beta-receptor system (Seloken 1 mg/ml Hässle, Göteborg, Sweden). Then 5–10 mg (mean, 7.5 ± 3.5 mg) of verapamil (Isoptin, 2.5 mg/ml Knoll, Germany) were given as a bolus dose followed by a continuous infusion of 10–20 mg/hr (mean, 14.3 ± 5.3 mg/hr) during the test. The doses were empirically chosen. The effect of the beta-blocker appeared immediately after the first injection of 5 mg of metoprolol, with a reduction in heart rate and blood pressure. Increasing the metoprolol to more than 20 mg did not have any obvious additional effect. Verapamil was then given in normal human doses as a bolus, followed by a continuous intravenous infusion. The combination of these two drugs causes biventricular depression, and to compensate for the right sided depression, a centrifugal pump was used to bypass the right ventricle. The line into the pump was a two-stage catheter inserted into the right atrial appendage and inferior caval vein, draining both the upper and the lower caval veins. The outflow line was inserted into the pulmonary trunk (Figure 1). To avoid competitive flow from the right ventricle and through the pulmonary valve, a tourniquet was tightened around the pulmonary artery, with the outflow cannula inside during the tests. Therefore, the entire minute volume was restricted to the centrifugal pump.

Figure 1
Figure 1:
A schematic view of the right ventricular bypass used in the heart failure model. The inflow cannula of the centrifugal pump was inserted by means of the upper caval vein into the right atrium, ending in the lower caval vein, and with the outflow cannula inserted into the pulmonary artery.

A special centrifugal pump was chosen (Rota-Flow, Jostra GmbH, Hechingen, Germany) that has the advantage that any flow rate can be chosen and fixed regardless of the afterload. The rationale was that substitution of a pump for the failing right ventricle would cause volume overload in the left ventricle and backward congestion. As a part of this congestion, the left atrial pressure was expected to increase. The goal was to achieve a left atrial pressure above 20 mm Hg, a value often observed in acute pulmonary edema.

Because the combination of beta-blocker and verapamil is known to produce arteriovenous blockade, a VVI pacemaker was used to control the heart rhythm. The minute volume was measured using two different methods: in the healthy animals, with a Transonic transducer on the pulmonary artery, and in the failure model, through the centrifugal pump. A comparison of the two systems was performed in a loop.

The loop consisted of a 12 mm tube starting at the outflow and ending at the inflow cannula of the centrifugal pump. The mid-part of the tube, however, was replaced by a thin walled polyurethane tube with an internal diameter of 17–18 mm to mimic the pulmonary trunk where the Transonic transducer was placed. The loop was filled with blood, and parallel measurements with different volume flows were performed. The measured differences were small (< 5%), and corrections were, therefore, omitted in the calculations (Figure 2).

Figure 2
Figure 2:
Comparison between the flow measured by the centrifugal pump and the Transonic transducer. There is excellent agreement between the measurements; difference in the measurements was <5% in the actual range, R2 = 0.99 and y = 0.0179 + 0.922.

Calculations of Heart Work

The left ventricular heart work is traditionally calculated as the work performed in 1 minute (left ventricular minute work –LVMW) and is expressed in kg × m/min and calculated according to the equation 18,19:MATHwhere MAP is the mean arterial pressure in mm Hg, LAP is the mean left atrial pressure in mm Hg, CO is cardiac output in L/min, and 0.0136 is the converting factor for mm Hg. According to the SI system, this is more correctly expressed in newton-meters or joules:MATH where 9.81 is the gravitational constant.

The mean hydraulic performance of the left ventricle in a minute can be expressed as left ventricular hydraulic performance (LVHP) in watts according to the following equation:MATHwhere 60 represents the seconds in a minute.

Example: The hydraulic performance of the left ventricle with a mean pressure gradient of 100 mm Hg (MAP − LAP) and a cardiac output of 5 L/min is:MATH

Data Collection and Statistics

The hemodynamic values were collected after instrumentation of the animals at baseline after a stabilization period of 15 minutes and a minimum of 10 minutes after induced failure. Hemodynamics were defined as stable when the LAP remained within ± 2 mm during 3 minutes of measurement. Data were collected digitally at 200 Hz, and mean values were calculated every 5 seconds. After the stabilizing period, mean values over a period of 3 minutes at baseline and in failure were collected. The data are expressed as mean values with one standard deviation. To compare the difference between the two conditions, a paired Student’s t-test was applied and regression analysis was done with Pearson’s correlation method. A p value < 0.05 was considered statistically significant, and the analyses were performed with commercial software (Microsoft Excel and Lotus 1–2-3).


The measured hemodynamic results are listed in Table 1. We compared the values at baseline with those after induced heart failure. There was no significant change in heart rate (96 ± 23 to 111 ± 21 beats/min;p = 0.24), or significant change in cardiac output: from 4.0 ± 0.8 compared with 3.5 ± 0.5 L/min in failure (p = 0.25). The LAP rose from 5 ± 3 to 25 ± 4 mm Hg (p < 0.001) (Figure 3), and the mean diastolic LV pressure rose from 2 ± 3 to 17 ± 7 mm Hg (p = 0.003). Mean pressure in the pulmonary artery rose from 17 ± 7 to 33 ± 5 mm Hg (p < 0.001), despite a nonsignificant decrease in pulmonary vascular resistance from 242 (± 81) to 172 (± 76) dynes × sec × cm−5 (p = 0.065). Table 2 displays the left ventricular hydraulic performance of each animal, which was reduced from 0.82 ± 0.27 to 0.43 ± 0.13 watts, whereas the corresponding left ventricular minute work was reduced from 5.0 ± 1.7 kg × m/min at baseline to 2.6 ± 0.8 kg × m/min in the failure condition (p = 0.001 for both).

Table 1
Table 1:
Comparison of Hemodynamic Measurements at Baseline and in Failure
Figure 3
Figure 3:
Relationship between left atrial pressure and minute volume. The lower cluster shows the individual healthy animals (squares), and the upper cluster (triangles) shows the animals after induction of heart failure. Basal mean minute volume, 4 ± 0.8; in failure, 3.5 ± 0.4 L/min. Basal left atrial pressure,5 ± 3; in failure, 25 ± 4 mm Hg. The lines show the change for the individual animals.
Table 2
Table 2:
Left Ventricular Hydraulic Performance of Each Animal


We have developed a model of isolated left ventricular heart failure in calves to test a left ventricular support system. The basic principle is to induce biventricular failure with drugs and a pacemaker and to bypass the right ventricle to achieve volume overload of the left ventricle. We have chosen two different drugs, a calcium antagonist and a beta-blocker, in combination with a VVI pacemaker. Other agents with a negative inotropic effect could also be used. Esmolol, a short acting beta-blocker, has been used in beating heart surgery to reduce the heart rate and make the ventricles more flaccid. 15,16

The pacemaker was primarily used to counteract the arteriovenous blockade that occurred in most of the animals. In some cases, the tube to the centrifugal pump collapsed due to negative pressure, preventing us from achieving a LAP equal to or above 20 mm Hg only by increasing the volume flow. In these cases, rapid ventricular pacing enabled us to obtain the desired increase in LAP. However, a systematic assessment of the contribution of the pacemaker to the rise in LAP was not done. Despite regular VVI pacing, the intraventricular pressure curves became more chaotic and some showed a tendency toward an alternans phenomenon. 20

Heart work at baseline and in failure was expressed as hydraulic performance in watts. A comparison of the heart work as stroke work (g × m/beat) would entail a comparison of the stroke work in sinus rhythm with paced beats under a more irregular LV pressure in failure. The pressure curves varied from beat to beat and probably included nonejecting ventricular contractions. Therefore, it was not considered meaningful to use the left ventricular end diastolic pressure (LVEDP) in failure compared with baseline in calculation of heart work. However, the rise in mean left ventricular diastolic pressure from 1.9 to 17 mm Hg (p = 0.003) may indicate an increased filling pressure and, in one animal with a stable sinus rhythm in failure, a clear increase in LVDP was observed (Figure 4).

Figure 4
Figure 4:
Left ventricular pressure tracing in animal 1. The left side of the diagram shows the pressure at baseline, and the right side shows the increased left ventricular end diastolic pressure in failure.

We did not measure the size of the left ventricle during the procedure, but increased filling pressures will entail an increase in the left ventricular end diastolic volume (LVEDV) as a compensation for the drug-induced negative inotropic effect by means of the Frank-Starling mechanism.

The left ventricular hydraulic performance describes the total external work of the left ventricle regardless of the rhythm and is an objective description of the power of the heart. Hydraulic performance in watts is also commonly used to describe the power of a mechanical cardiac assist device, and applying this concept to heart work makes it feasible to estimate the contribution of an assist device to the circulation.

Despite the small and nonsignificant reduction in cardiac output, calculations of left ventricular mechanical work revealed a substantial reduction in hydraulic performance from 0.82 at baseline to 0.43 watts in failure, mainly due to the reduced mean arterial and increased left atrial pressures (Table 1).

Our goal of creating a left ventricular heart failure model with an LAP above 20 mm Hg was accomplished. The increased pressure in the pulmonary artery distal to the tourniquet indicates a backward pulmonary congestion, as the pulmonary vascular resistance remained virtually unchanged.

The tourniquet around the pulmonary trunk allowed no outlet for the right ventricle, which explains the increased pressure developed in the right ventricle. Hence, the pressures are generated by a nonejecting right ventricle and, therefore, represent an isovolumic pressure generation, provided that no tricuspid regurgitation occurred. It is difficult to judge how long an animal can be kept stable with this model, but a longer period is conceivable, because the procedure causes no damage to the myocardium except for the instrumentation of the heart.

In conclusion, with negative inotropic agents, a circulatory pump, and pacemaker, it is possible to create a left ventricular failure model useful for performing short-term hemodynamic studies. The model has several advantages: ischemic procedures to bring about the heart failure are avoided, a pure left ventricular failure is accomplished, and the right ventricular bypass is adjustable, making it possible to obtain the desired level of filling pressure of the left ventricle. In this setting, it would be possible to perform acute hemodynamic studies of a cardiac assist device. As expected, long-term effect of a support device will always be based on its acute hemodynamic effect, and this is probably best examined before an adaptation to the support system has developed.


1. Butler KC, Dow JJ, Litwak P, Kormos RL, Borovetz HS: Development of the Nimbus/University of Pittsburgh innovative ventricular assist system. Ann Thorac Surg 68: 790–794, 1999.
2. Fossum TW, Morley D, Benkowski R, et al: Chronic survival of calves implanted with the DeBakey ventricular assist device. Artif Organs 23: 802–806, 1999.
3. Marlinski E, Jacobs G, Deirmengian C, Jarvik R: Durability testing of components for the Jarvik 2000 completely implantable axial flow left ventricular assist device. ASAIO J 44: M741–M744, 1998.
4. Rakhorst G, van der Meer J, Kik C, et al: Biological evaluation of mechanical circulatory support systems in calves. Int J Artif Organs 19: 472–476, 1996.
5. Harasaki H, Fukamachi K, Benavides M, Manos J, Wika K, Massiello A: A comprehensive hematologic study in calves with total artificial hearts. ASAIO J 41: M266–M271, 1995.
6. Reitan O, Ohlin H, Peterzén B, Granfeldt H, Steen S, Emanuelsson H: Initial tests with a new cardiac assist device. ASAIO J 45: 317–321, 1999.
7. Reitan O, Sternby J, Öhlin H: Hydrodynamic properties of a new percutaneous intra aortic axial flow pump. ASAIO J 46: 323–329, 2000.
8. Nakazawa T, Shiono M, Inoue T, Orime Y, Shindo S, Sezai Y: Preliminary experimental study about the feasibility of combining pulsatile cardiopulmonary support system and a membrane oxygenator. Int J Artif Organs 19: 597–604, 1996.
9. Kovacs SG, Reynolds DG, McKeown PP, et al: A magnetically actuated left ventricular assist device. ASAIO J 38: 38–46, 1992.
10. Shiiya N, Zelinsky R, Deleuze PH, Loisance DY: Effects of Hemopump support on left ventricular unloading and coronary blood flow. ASAIO Trans 37: M361–M362, 1991.
11. Risoe C, Hall C, Smiseth OA: Blood volume changes in liver and spleen during cardiogenic shock in dogs. Am J Physiol 261 (Pt 2): H1763–H1768, 1991.
12. Grund F, Sommerschild HT, Lyberg T, Kirkeboen KA, Ilebekk A: Microembolization in pigs: Effects on coronary blood flow and myocardial ischemic tolerance. Am J Physiol 277: H533–H542, 1999.
13. Scholz KH, Hering JP, Schroder T, et al: Protective effects of the Hemopump left ventricular assist device in experimental cardiogenic shock. Eur J Cardiothorac Surg 6: 209–214, 1992.
14. Schmidt C, Schwendenwein I, Wieselthaler G, et al: Pharmacologically induced heart failure for the evaluation of circulatory assistance. Artif Organs 20: 685–688, 1996.
15. Casimir-Ahn H, Lonn U, Peterzen B: Clinical use of the Hemopump cardiac assist system for circulatory support. Ann Thorac Surg 59 (Suppl 2): S39–S45, 1995.
16. Lonn U, Peterzen B, Granfeldt H, Casimir-Ahn H: Coronary artery operation with support of the Hemopump cardiac assist system. Ann Thorac Surg 58: 519–522, 1994; discussion 523.
17. Braunwald E: Heart Disease. A Textbook of Cardiavascular Medicine, 5th edition. Philadelphia: Saunders, 1997, pp. 67, 1411.
18. Swanton RH: Cardiology. 3rd edition. Oxford: Blackwell Scientific Publications, 1994, pp. 561–562.
19. Vij D, Babcock R, Magilligan DJ: A simplified concept of complete physiological monitoring of the critically ill patient. Heart Lung 10: 75–82, 1981.
20. Baller D, Hoeft A, Korb H, Wolpers HG, Zipfel J, Hellige G: Basic physiological studies on cardiac pacing with special reference to the optimal mode and rate after cardiac surgery. Thorac Cardiovasc Surg 29: 168–173, 1981.
Copyright © 2002 by the American Society for Artificial Internal Organs