Secondary Logo

Journal Logo

Adult Circulatory Support

Hemodynamic and Metabolic Effects of Para- versus Intraaortic Counterpulsatile Circulation Supports

Lu, Pong-Jeu*†; Lin, Pao-Yen; Yang, Chi-Fu Jeffrey§; Hung, Chun-Hao*; Chan, Ming-Yao*†; Hsu, Tzu-Cheng*†

Author Information
doi: 10.1097/MAT.0b013e3181fcbc7d
  • Free


The intraaortic balloon pump (IABP) is the most widely used mechanical circulatory support device with well-documented beneficial hemodynamic and metabolic effects.1–5 However, it is rarely used long term due to several drawbacks, including immobility, limb ischemia (it is especially not tolerated in patients with preexisting peripheral vascular disease), and limited effectiveness in the face of severe cardiogenic shock and low aortic pressures.1,3,5,6

There have been continuing efforts to design a counterpulsation device more suitable for long-term use7–15 given the potential for chronic contraction unloading to induce reverse modeling and myocardial recovery and offer benefit to patients with early- or intermediate-stage heart failure.13,16–18 Preliminarily, there is growing evidence to suggest that a counterpulsation device could be safer, simpler to implant, and far less costly, than the typical left ventricular assist devices (LVADs) available today, while providing greater hemodynamic benefit than the IABP.7–15

Current counterpulsation research focuses on hemodynamic indices, such as the reduction of end-diastolic aortic pressure, to assess efficacy.1–3 However, emphasizing a few hemodynamic indices in isolation without examining other key issues—such as coronary blood regurgitation (coronary blood steal),14 myocardial oxygen demand/supply calculations, and inflation/deflation timing control—can lead to bias in performance interpretation.

The purposes of this study were 1) to compare the performance of the IABP to a novel paraaortic blood pump (PABP)11 using traditional performance indices and 2) to introduce more physiologic performance indices and a timing optimization scheme that heretofore have not been applied to counterpulsation research.

Materials and Methods

Counterpulsation Device

The present PABP device, which is made from polyurethane, is depicted in Figure 1. It consists of a T-manifold connected seamlessly on top of an oval-shaped bladder. The T-manifold conduit and bladder assembly is housed within a semirigid outer shell on which a deairing stub and a drive line are installed. The thin (3 mm diameter) drive line accepts pneumatically driven gases percutaneously, provided either by bedside console or by wearable driver. The present T-manifold is valveless and has a sharp-edged, compliance-matching graft-host juncture design.11 When anastomosed to the aorta with appropriate oversizing, this inserted manifold can snuggly be embraced by the lumen of the aortic wall. The streamlined, sharp-edged conduit ends (∼0.1 mm in thickness) can dilate and recoil in concert with the aortic pressure, potentially reducing the possibility of graft complications associated with the pump inflow and outflow tract design.19

Figure 1.
Figure 1.:
Paraaortic blood pump (PABP). A: A sketch showing the components and the insertion in the descending aorta. B: A photograph of a PABP prototype (40 ml stroke volume) showing the semirigid outer shell.

The PABP was implanted in the descending aorta by left thoracotomy; implantation was performed without cardiopulmonary support. During pump insertion, the proximal and distal ends of the device implant site were temporarily cross-clamped. Only simple suture was needed to secure the PABP after intraaortic T-manifold embedment followed by the subsequent release of the cross-clamp. The cross-clamped period, hence, was generally maintained for <3 minutes.

Animal Model and Instrumentation

The acute studies were performed on eight domestic Lanyu mini pigs, weighing from 60 to 80 kg (mean: 67.4 ± 2.7 kg). This study was approved by the Experimental Animal Committee of National Cheng Kung University Medical Center. The animals received humane care in compliance with the “Guide for the Care and Use of Laboratory Animals” (NIH publication no. 86-23, revised 1996). These pigs were randomly divided into two study groups: IABP group (n = 4) and PABP group (n = 4). Animals were premedicated by subcutaneous injection of the mixture of 6 mg/kg Zoletil plus 1.5 mg/kg Rompan and 0.02 mg/kg Atropine. Anesthesia was maintained using isoflurane in oxygen after endotracheal intubation. Left thoracotomy was performed at the fifth intercostal space with the fifth and sixth ribs removed to obtain chest access and to create space for the installation of measurement probes. Pericardium was incised, and a cradle was made to expose the heart. Cannulations were made in the left carotid artery and jugular vein to measure ascending aortic pressure (AoP) and central venous pressure (CVP), respectively; the right atrium apex to measure the right ventricular pressure (RVP); the left atrium apex to measure the left ventricular pressure (LVP). The left anterior descending (LAD) artery was dissected free for the coronary flow rate (CoF) measurement using a Transonic perivascular probe MA 2.5PSS (Transonic Systems Inc., Ithaca, NY). Both aortic (AoF) and pulmonary arterial (PAF) flow rates were measured using the Transonic probes MA20PAX. Additional pressure and flow rate measurements (for the PABP group only) were performed in the up- and downstream sites around the anastomosed T-manifold.

For the IABP counterpulsatile study, Datascope Fidelity 8Fr. 40cc IAB catheters (Maquet Cardiovascular LLC, Wayne, NJ) were inserted through the left femoral artery following the standard deployment protocol. IABPs were positioned in the descending aortic area. PABP placement was in the same location as IABP placement. Pumping action was initiated using the Datascope System 98XT console (Maquet Cardiovascular LLC, Wayne, NJ), for both IABP and PABP experiments. Except for occasional monitoring line flushing, no anticoagulant was administered.

Data Reduction and Statistical Analysis

All blood pressure measurements were obtained using a pressure transducer system DPT-248 (Utah Medical Products Inc., West Midvale, UT). We transmitted all acquired pressure and flow rate data into a data acquisition (DAQ) card PCI-6229 (National Instrument, Austin, TX) driven by Labview (National Instrument, Austin, TX) software system. Postprocessing filter software available in MATLAB (MathWorks, Natric, MA) was adopted to precondition the data.

Throughout the present comparison study, counterpulsatile support enforced either by IABP or by PABP was carried out in a 1:3 (1 pumping assistance/3 heart beats) fashion. The R-wave recognized on the electrocardiogram (ECG) was taken as the initiation time of each beat cycle. For every three heart beats, diastolic augmentation and systolic unloading occurred over the diastolic and systolic phases pertaining to the first and the second beats, respectively; whereas the third beat was taken as the unassisted control beat. Counterpulsatile support effectiveness was thus evaluated using data taken from the first two assisted beats and then compared against those obtained from the third unassisted control beat. Data were phase averaged over 60 beats (20 periods) after hemodynamic states stabilized with periodicity established. All time scales of the figures were normalized into 1 second per beat, so as to ease the presentation of comparison.

Statistical analysis was performed using SPSS for Windows (Version 11.0, SPSS, Inc., Chicago, IL), and Mann-Whitney U test was adopted for comparisons between the IABP and PABP groups; Student's t-test for the paired data obtained under assisted and control conditions within the same device group.

Counterpulsation Performance Indices

End-Diastolic and Peak Diastolic Aortic Pressures.

Traditionally, decreases in end-diastolic aortic pressure are taken as a measure of systolic unloading and considered beneficial.1,2 Higher peak diastolic aortic pressure is interpreted as a stronger driving potential for myocardial perfusion augmentation.1,2

Tension Time Index.

Tension time index (TTI) is an approximation of the effort exerted by the left ventricle (LV) in ejection, indicating myocardial oxygen consumption, which is defined as a time integration of the LVP in systole1,20,21:

Endocardial Viability Ratio.

Oxygen supply has been approximated to be in proportion to coronary perfusion and, hence, depends on the mean diastolic aortic pressure in the aorta. As a result, the concept of oxygen supply to consumption ratio in the myocardium is given the name endocardial viability ratio (EVR), which is defined by:

The most critical index for counterpulsation optimization is thought to be the EVR.1,21,22 However, both TTI and EVR are derived from general approximations, based on the availability of LV and aortic pressures, rather than from rigorous physiologic energy considerations. Further, the numerator of Equation 2 ignores the systolic contribution to oxygen supply and the possible coronary regurgitation (“blood steal”) penalty caused by a strong systolic unloading.

LV Stroke Work.

Stroke work (SW) represents the work delivered to the ejected blood stream per heart beat23 and is defined as follows:

in which dV denotes differential volume change and can be related to ventricular ejection flow rate AoF by dV = −AoFdτ. The subscript “loop” indicates a line integral of the pressure-volume loop (P-V loop) of the LV.

Note, the aforementioned index TTI can be viewed as a quantity proportional to SW produced under a constant ejection volume flow rate—this is obviously not true in any realistic blood ejection motion, and therefore, we propose that LV SW is a more physiologic approximation of myocardial oxygen demand than TTI.

For our present interest of LV SW, the systolic left ventricular volume (LVV) was defined as follows:

where Ved indicates the end-diastolic LVV. A P-V loop plot can thus be generated using this integrated LVV and the measured LVP. Stroke volume, cardiac output enhancement, and LV SW can be deduced accordingly using this P-V loop.

Coronary Perfusion.

In general, epicardial CoF represents a gross approximation of the oxygen and nutrition that is supplied to the myocardium as a whole. Thus, in this study, the coronary perfusion performance was assessed using the time integration of the LAD flow rate in an entire beat cycle, as defined by:

This method of calculation has its advantages over traditional approaches that rely on aortic pressure measurements, because it is able to account for the regurgitated CoF that may occur near systole onset due to the unloading-induced reduction of the end-diastolic aortic pressure generated by the counterpulsation device.

Revised Endocardial Viability Ratio.

In our experiments, sufficient relevant pressure and flow measurements were available, hence allowing the calculation of coronary perfusion and SW and allowing calculation of a more physiologically compatible index that may better describe the oxygen supply and demand ratio of the heart. This revised EVR is termed EVRnew and is given by:

Timing of IABP and PABP Deflation

A conventional IABP timing1,2—where inflation of balloon begins at the dicrotic notch and deflation of the balloon occurs immediately before ventricular systole—was used. For PABP actuation, we adopted the same inflation timing control as that of the IABP. As for our PABP deflation timing control, however, the deflation timing was delayed one count (25 milliseconds) relative to that of the IABP unloading control. This was because in our preliminary testing, when the standard IABP deflation timing was applied to the PABP, excessively low end-diastolic pressure and regurgitated CoF were observed. Therefore, to reasonably compare the counterpulsation effectiveness, we delayed the PABP deflation timing by one count, so as to make the end-diastolic pressures achieved by IABP and PABP deflation comparable.

We also experimented with different deflation timing possibilities in one pig model, using deflation timing options at 1) one count (25 milliseconds) preceding the R-wave, 2) the R-wave, 3) one count (25 milliseconds) after the R-wave, and 4) two counts (50 milliseconds) after the R-wave.


Classical Counterpulsation Performance Indices

Figure 2 shows representative hemodynamic waveforms of a pig supported by IABP or PABP counterpulsation. Counterpulsation efficacy can be evaluated by comparing the data pertaining to assisted and unassisted beats. Classical performance indices are summarized in Table 1. The classical quantification of systolic unloading (decrease of end-diastolic aortic pressure [AoPed] or peak systolic LV pressure [LVPpksys]) and the measure of diastolic augmentation by the increase of the peak diastolic aortic pressure (AoPpkdias) were demonstrated by both the IABP- and the PABP-assisted groups. It can be observed that the present PABP outperformed the IABP in counterpulsation efficacy, according to classical measures. It is observed that the decrease in AoPed contributed by IABP and PABP is −13.2% ± 9.9% vs. −34.0% ± 17.1%, p = 0.2; the decrease in LVPpksys is −8.50% ± 5.9% for IABP vs. −26.3% ± 14.0% for PABP, p = 0.05; and the increase in AoPpkdias is 27.5% ± 11.7% for IABP vs. 41.4% ± 3.3% for PABP, p = 0.03.

Figure 2.
Figure 2.:
Hemodynamic characteristics of representative porcine blood flow circulation under 1:3 counterpulsation support. A: IABP supported. B: PABP supported. AoP, aortic pressure measured at ascending aorta; LVP, left ventricular pressure; RVP, right ventricular pressure; PAF, pulmonary arterial flow rate; AoF, ascending aortic flow rate; CoF, coronary flowrate measured at left anterior descending coronary artery; ECG, electrocardiogram; PABP, paraaortic blood pump; IABP, intraaortic balloon pump.
Table 1
Table 1:
Comparison of Counterpulsation Performance between IABP and PABP Groups (Classical Indices)

Physiologic Counterpulsation Performance Indices

In this study, TTI, EVR, EVRnew, CoPerf, and LV SW were selected as the major metabolic counterpulsation indices that were computed postoperatively for both the IABP (n = 4) and PABP (n = 4) groups. Generally, both IABP assistance and PABP assistance improved physiologic indices (Table 2). However, the PABP-supported group had a greater reduction in LV SW than did the IABP-supported group (−6.2% ± 2.9% for IABP vs. −14.6% ± 7.7% for PABP), a finding that approached statistical significance (p = 0.08).

Table 2
Table 2:
Comparison of Counterpulsation Performance between IABP and PABP Groups (Physiologic Indices)

Both counterpulsation modalities were effective in increasing coronary perfusion and had periods of reversed coronary waveforms using the standard IABP inflation/deflation timing. However, when we adjusted the timing of the PABP control to allow for delayed pump deflation, reversed CoF could be reduced to a minimum (Figure 3).

Figure 3.
Figure 3.:
Effect of deflation timing on PABP-supported coronary perfusion. Inflation timing is set on the dicrotic notch, and zero-deflation timing is timed on the R-wave. AoP, aortic pressure measured in the ascending aorta; CoF, coronary flow rate measured at left anterior descending coronary artery; PABP, paraaortic blood pump.

Deflation Timing Optimization

A preliminary PABP deflation optimization was demonstrated for pig 8 with corresponding results in LV SW, TTI, EVR, EVRnew, and CoPerf, listed in Table 3. The dependence of coronary regurgitation (blood steal) on deflation timing is illustrated in Figure 3. It can be observed that early deflation may result in shortening of the diastolic perfusion period and increasing the reversed coronary blood flow before systole.

Table 3
Table 3:
Effect of Deflation Timing on PABP Counterpulsation Performance


Traditionally, counterpulsation performance is determined using classical hemodynamic indices based on the principles that during counterpulsation: 1) diastolic augmentation elevates the aortic pressure in diastole to enhance coronary perfusion and 2) systolic unloading provides the heart with a reduced afterload during ventricular contraction and ejection. Thus, typical hemodynamic indices used to determine counterpulsation efficacy include: 1) the end-diastolic aortic pressure reduction and/or peak systolic LVP reduction achieved (a measure of systolic unloading) and 2) the increase in peak diastolic aortic pressure (a measure of diastolic augmentation).1,2 In this study, we found a novel PABP to have an average beneficial effect on these hemodynamic indices 1.5–3.5 times greater than that of the IABP in an acute porcine model.

Many counterpulsation studies draw their conclusions primarily on these aforementioned pressure variations, which provide limited physiologic information. For example, large reductions in end-diastolic aortic pressure achieved by aortic unloading typically is seen as positive but can lead to retrograde CoF,14 where blood originally flowing to the heart through the coronary arteries is instead “sucked” back into the aorta.

To further assess performance on a more physiologic basis, we have incorporated the metabolic indices of TTI (a measure of myocardial oxygen demand) and EVR (the ratio between myocardial oxygen supply and demand) in our analysis. During the systolic unloading phases of both the IABP and PABP supports, the TTI decreased, suggesting that both devices decreased myocardial oxygen consumption. Both the IABP and PABP supports increased the EVR, a sign that the myocardial supply/demand equilibrium and coronary perfusion improved during both approaches.

However, because the TTI and EVR are derived solely on pressure measurements, and because EVR is based only on diastolic measurements in the evaluation of coronary perfusion, they are not precise physiologic approximations of myocardial oxygen demand and supply. Despite these limitations, previous studies still used TTI and EVR because flow measurements were unavailable. In our study, we were able to measure flow rates and pressures, which allowed us to quantify true coronary perfusion (during both systole and diastole) and ventricular effort in a more direct manner. In addition, we derived EVRnew, which is a closer portrait of the supply/demand condition for the assisted myocardium than EVR. This is because, as described in detail earlier in the text, coronary perfusion (used in EVRnew) is a better approximation of myocardial oxygen supply than the integration of the diastolic aortic pressures (used in EVR) and SW is a better approximation of oxygen demand than myocardial TTI.

The modification of these classical counterpulsation performance indices, hence, resulted in the proposition of using coronary perfusion (a measure of myocardial oxygen supply), left ventricular SW (a measure of myocardial oxygen demand), and EVRnew (the supply to demand ratio for the heart) as the main criteria for the evaluation of counterpulsation performance.

Based on these criteria, the IABP and PABP were found to be comparable and to have a beneficial effect—both decreased TTI and SW, and increased EVR and EVRnew. The main difference was, compared with the IABP, the PABP was able to achieve greater reduction in myocardial oxygen demand through greater SW reduction, a finding that approached statistical significance (p = 0.08).

Both the IABP and PABP caused periods of reversed CoF, despite the net CoF for both the IABP and PABP having increased due to diastolic augmentation. Previous studies, including ours, use a traditional timing approach for the PABP, which is similar to that used in an IABP device. This approach, however, may be suboptimal for PABP because it can lead to excessively low end-diastolic aortic pressure and coronary blood steal.

Counterpulsation effectiveness depends sensitively on the inflation/deflation timing selected for the control system of the driving console.1,2,21 Because the IABP is in the lumen of the aorta, its inflation and deflation must be timed, so that it does not occlude the aorta during systole. The PABP is nonocclusive, in that the pump is outside of the aorta; the deflation timing interval can thus extend into systole. This can allow for experimentation with a new optimal PABP inflation/deflation timing.

To our knowledge, there have been no previous studies of chronic counterpulsation (non-IABP) devices quantifying the effect of timing control on physiologic performance indices. Our preliminary PABP deflation optimization experiment (Table 3) demonstrated that the indices of TTI, SW, and EVRnew achieved optimal scores when the pump deflation was delayed one count into LV ejection (25 milliseconds after the R-wave); this is in contrast to the conventional IABP timing where deflation is usually set before the R-wave.

Coronary perfusion augmentation was found to be most sensitive to PABP deflation control—the percent perfusion augmentation (0.0% ± 3.6%) was lowest using a standard IABP deflation setting, and increased to as high as 15.5% ± 3.6% when deflation was two counts after LV ejection. The corresponding coronary waveforms (Figure 3) show that a two-count delay in the deflation control setting (50 milliseconds after R-wave) was able to almost completely eliminate coronary blood steal during systolic ejection. This demonstrates the inherent flexibility of PABP deflation timing with respect to coronary perfusion—which is that device-induced coronary regurgitation or blood steal can be avoided with a sensible manipulation of the deflation control. The present findings suggest that optimized PABP deflation, depending on how much coronary augmentation is desired, may course into the LV ejection phase without incurring the traditional performance degradation of ejection obstruction associated originally with IABP.

We have demonstrated that classical counterpulsation indices typically used in counterpulsation research, when interpreted in isolation can be inadequate. Future research of counterpulsation may benefit from a combined examination of 1) traditional hemodynamic indices, 2) traditional oxygen supply/demand indices such as TTI and EVR, 3) more physiologic approaches to the evaluation of myocardial oxygen supply and demand, using coronary perfusion, SW, and EVRnew, 4) examination of coronary waveforms to measure coronary blood steal, and 5) timing control, with future research studying timing control optimization under selected objective functions in different clinical settings including those of heart failure and/or stenosed coronary arteries.

Note that the number of pigs used in this study was small (n = 8). Further animal studies using a bovine model are currently underway to evaluate PABP counterpulsation. A heart failure model and chronic survival tests will be included in those calve investigations. Future studies should also aim to construct a pump that can be implanted using minimally invasive technique and to design a thrombo-resistant PABP that can enforce long-term counterpulsation for patients with moderate-to-severe heart failure.


In summary, a comparison between the IABP and a novel PABP counterpulsatile circulation support was conducted using acute porcine experiments. Evaluation in terms of traditional hemodynamic indices and metabolic indices of TTI, EVR, coronary perfusion, and left ventricular SW was performed. The PABP counterpulsatile performance, when enforced by way of the standard IABP inflation/deflation control guideline, is comparable, if not superior, to that of the IABP; however, there will still be a period of penalized coronary blood regurgitation, which can be further minimized by optimizing the PABP deflation timing control. This possible elimination of coronary blood steal is not attainable by intravascular IABP counterpulsation. Therefore, PABP counterpulsation may be superior to that of IABP provided that this delayed deflation control is optimized.


Supported partially by National Science Council (Grant NSC 98-2221-E-006-030) of Taiwan and the University Excellence Fund of National Cheng Kung University, and by the Howard Hughes Medical Institute Medical Student Fellowship (C.-F.J.Y.). The authors thank Dr. Willard Daggett for reviewing the manuscript and for very useful discussions on counterpulsation.


1. Papaioannou TG, Stefanadis C: Basic principles of the intraaortic balloon pump and mechanisms affecting its performance. ASAIO J 51: 296–300, 2005.
2. Bolooki H (ed): Clinical Application of the Intra-Aortic Balloon Pump, 3rd ed. New York, Futura, 1998.
3. Nanas JN, Moulopoulos SD: Counterpulsation: Historical background, technical improvements, hemodynamic and metabolic effects. Cardiology 84: 156–167, 1994.
4. Ferguson JJ 3rd, Cohen M, Freedman RJ Jr, et al: The current practice of intra-aortic balloon counterpulsation: Results from the benchmark registry. J Am Coll Cardiol 38: 1456–1462, 2001.
5. Dunkman WB, Leinbach RC, Buckley MJ, et al: Clinical and hemodynamic results of intraaortic balloon pumping and surgery for cardiogenic shock. Circulation 46: 465–477, 1972.
6. Arafa OE, Pedersen TH, Svennevig JL, et al: Vascular complications of the intraaortic balloon pump in patients undergoing open heart operations: 15-year experience. Ann Thorac Surg 67: 645–651, 1999.
7. Utoh J: A history of the valveless paraaortic counterpulsation device. J Thorac Cardiovasc Surg 112: 850–851, 1996.
8. Legget ME, Peters WS, Milsom FP, et al: Extra-aortic balloon counterpulsation: An intraoperative feasibility study. Circulation 112 (suppl I): I-26–I-31, 2005.
9. Nanas SN, Nanas JN, Charitos CE, et al: High stroke volume para-aortic counterpulsation device versus centrifugal pump in cardiogenic shock: Experimental study. World J Surg 21: 318–322, 1997.
10. Jeevanandam V: The quest for permanent ventricular assistance: The role of aortic counterpulsation, in ASAIO 50th Anniversary Highlights, Awards & Related Articles. ASAIO J 50: xxxvii–xlii, 2004.
11. Lu PJ, Lin PY, Chern L, et al: Dual-pulsation bi-ventricular assist device. US Pat Appl Publ, US20080300447A1, 2008.
12. Terrovitis JV, Charitos CE, Tsolakis EJ, et al: Superior performance of a para-aortic counterpulsation device compared to the intraaortic balloon pump. World J Surg 27: 1311–1316, 2003.
13. Koenig SC, Spence PA, Pantalos GM, et al: Development and early testing of a simple subcutaneous counterpulsation device. ASAIO J 52: 362–367, 2006.
14. Koenig SC, Litwak KN, Giridharan GA, et al: Acute hemodynamic efficacy of a 32-ml subcutaneous counterpulsation device in a calf model of diminished cardiac function. ASAIO J 54: 578–584, 2008.
15. Astra LI, Stephenson LW: Skeletal muscle as a myocardial substitute. Proc Soc Exp Biol Med 224: 133–140, 2000.
16. Heerdt PM, Holmes JW, Cai B, et al: Chronic unloading by left ventricular assist device reverses contractile dysfunction and alters gene expression in end-stage heart failure. Circulation 102: 2713–2719, 2000.
17. Nakatani S, McCarthy PM, Kottke-Marchant K, et al: Left ventricular echocardiographic and histologic changes: Impact of chronic unloading by an implantable ventricular assist device. J Am Coll Cardiol 15: 894–901, 1996.
18. Birks EJ, Tansley PD, Hardy J, et al: Left ventricular assist device and drug therapy for reversal of heart failure. N Engl J Med 355: 1873–1884, 2006.
19. Houël R, Moczar M, Ginat M, Loisance D: Pseudointima in inflow and outflow conduits of a left ventricular assist system: Possible role in clinical outcome. ASAIO J 47: 275–281, 2001.
20. Sarnoff SJ, Braunwald E, Welch GH Jr, et al: Hemodynamic determinants of oxygen consumption of the heart with special reference to the tension-time index. Am J Physiol 192: 148–156, 1958.
21. Zelano JA, Li JK, Welkowitz W: A closed-loop control scheme for intraaortic balloon pumping. IEEE Trans Biomed Eng 37: 182–192, 1990.
22. Barnea O, Moore TW, Dubin SE, Jaron D: Cardiac energy considerations during intraaortic balloon pumping. IEEE Trans Biomed Eng 37: 170–181, 1990.
23. Sagawa K, Maughan L, Suga H, Sunagawa K: Cardiac Contraction and the Pressure-Volume Relationship. New York, Oxford University Press, Inc, 1988.
Copyright © 2011 by the American Society for Artificial Internal Organs