The incidence of heart failure (HF) is increasing worldwide, with more than one million new cases diagnosed annually. Even with well-managed care, medical therapy is insufficient to sustain an acceptable level of cardiac function for most patients with severe HF. Cardiac transplantation that can provide an optimal treatment, however, is significantly restricted by recent decline in organ donation.1 Cardiac mechanical support therefore is expected as a bridge to recovery for advanced HF.2,3
Para-aortic counterpulsation device (PACD) is a pneumatically driven, polyurethane chamber with a valveless orifice connected to the ascending aorta by a Dacron vascular graft. With principle of counterpulsation augmentation, it can significantly unload the left ventricle (LV) and provide an effective counterpulsation wave, even at very low aortic pressure.4–8
However, since the first report by Nose in 1963,9 PACD has only presented in several animal experiments and limited clinical applications.10 Although Giridharan et al.11 tested the counterpulsation device in a computer mock circulation and suggested that a early filling late ejection algorithm might provide optimal hemodynamic support, it is still unclear what algorithm is required to provide optimal circulatory assistance in vivo.
Materials and Methods
After initial physical examination and a 5 to 7 day quarantine period, six healthy sheep (45 ± 3 kg) were used in this study. All animals received human care and experimental procedures were carried out strictly in compliance with the 1996 “Guide for the Care and Use of Laboratory Animals” published by the National Institutes of Health.
The animals were anesthetized with ketamine (10 mg/kg) and midazolam (0.2 mg/kg) intramuscularly. After intravenous injection of succinylcholine (2 mg/kg) for muscle relaxation, the animals were ventilated at a rate and depth appropriate to weight (e.g., 800 ml tidal volume at 18 breaths per minute for a 40 kg sheep). Anesthesia was maintained by an inhalation of 1–3% enflurane and an intravenous injection of propofol 4 mg/kg/h. The electrocardiogram (ECG) and peripheral oxygen saturations were monitored.
A 7F Swan-Ganz catheter was advanced into the pulmonary artery for the determination of cardiac output (CO), pulmonary arterial wedge pressure, and central venous pressure via right external jugular access. Three pressure catheters were placed in left atrium (LA), LV, and aorta, respectively, to monitor pressure. To measure flow volume, an ultrasonic flow probe (Transonic Systems, Inc., Ithaca, NY) was set around the left mean coronary artery (LM) and a Doppler ultrasound probe was placed on left carotid artery. Blood collected from aorta and coronary sinus was used for blood gas analysis. Hemodynamic data were recorded by MP150 system (Biopac Systems, Inc., Goleta, CA) with output to computer at a sampling rate of 1 kHz, by using AcqKnowledge software.
A bolus of heparin was administered (2 mg/kg) for anticoagulation. After median sternotomy, a PACD with 40 ml stroke volume was implanted, as shown in Figure 1.
Acute LV failure was induced by multiple ligation of coronary arterial branches and continuous intravenous administration of propranolol, as described previously by Terrovitis et al.12 After successful establishment of the model (LV end-diastolic pressure above 15 mm Hg), the number of ligated coronary artery and dose of propranolol continued to be adjusted according to the following standards: CO was approximately 2.0 L/min and aortic systolic pressure was 70–80 mm Hg, to avoid interference from different hemodynamic conditions for experimental results. To prevent ventricular arrhythmias, a bolus dose of lignocaine was given intravenously (1 mg/kg) before the ligation, and then was maintained at 1 mg/min/kg by intravenous dripping.
The device was driven by an intra-aortic balloon pumping (IABP) pump with ECG trigger. In the present study, 1:2 support (1 counterpulsation for 2 cardiac cycles) was used. The deflation (blood filling) was triggered by pre-R-wave or R-wave in ECG, and the inflation (ejection) was triggered by dicrotic notch of aortic pressure waveform or postdicrotic notch. So there were four triggering algorithms as follows: pre-R-wave-deflation and dicrotic notch-inflation (PD), pre-R-wave-deflation and postdicrotic notch-inflation (PP), R-wave-deflation and dicrotic notch-inflation (RD), and R-wave-deflation and postdicrotic notch-inflation (RP). For each sheep, hemodynamic data were accumulated during a minimum of six assisted cycles after a period of at least 1 minute of assist to allow the circulation to reach equilibrium.
Indices of Support Effectiveness
Hemodynamic support effectiveness was assessed through mean aortic diastolic pressure (MADP), coronary artery blood flow and peripheral artery blood flow, the endocardial viability ratio (EVR; also known as diastolic augmentation index, a simplified version of the ratio of the diastolic pressure time index to the tension time index).13 The LM flow (QLM) and carotid artery flow (QCA) were measured. The EVR was calculated from the following formula:
The pulmonary artery flow/LA pressure ratio (QPA/LAP) is an index of left heart pump function.14 The CO is equal to the QPA, and thus QPA/LAP = CO/LAP. External LV work (ELVW; J/min) is calculated as follows15:
Left ventricular myocardial oxygen consumption (LVVO2), as an important indicator of cardiac work, is calculated as follows16:
Cardiac mechanical efficiency (CME) is calculated as ELVW/LVVO2.
All offline measurements were made by investigators blinded to models. Results are presented as mean ± SD. Statistical analysis was performed by Statistical Product and Service Solutions (SPSS) 13.0 software package (SPSS Inc., Chicago, IL). To eliminate interference from different circulatory conditions of different animals, paired Student t-test was used for self-comparison between different PACD assistant models and nonassist, and percent changes of all parameters at different assistant model were compared through one-way analysis of variance with Bonferroni post hoc correction.16,18,19p < 0.05 was regarded as significant statistical difference. Linear correlation analysis using a Pearson’s test was performed to assess relationship among all parameters.
Of the eight sheep studied, two were excluded from the further experiments because of an intractable ventricular fibrillation after ligation of coronary artery and an excessively severe and out of range HF, six were determined ultimately for the subsequent procedures.
The photographs of electrocardiogram, aortic and LV pressure, LM flow, and QCA waveforms are displayed in Figure 2. The following analysis was used to assess hemodynamic benefits.
Table 1 shows hemodynamic and LV functional data and percent changes of these parameters are compared in Figure 3. Compared with nonassist, the mean aortic systolic pressure (MASP) significantly decreased and MADP significantly increased at all PACD assistant models (p < 0.005). There were bigger decrements in MASP at RD (−14 ± 3%, 95% CI: −17%, 11%) and RP (−11 ± 2%, 95% CI: −13%, −8%) models than that at PD (−4 ± 2%, 95% CI: −6%, 2%) and PP (−4 ± 2%, 95% CI: −6%, 2%; p < 0.001) models, but no differences between PD and PP models (p = 1) and between RD and RP models (p = 0.199). The increments in MADP were more at RD (11 ± 2%, 95% CI: 8%, 13%) than that at PD (6 ± 2%, 95% CI: 5%, 8%; p = 0.01) and PP models (3 ± 2%, 95% CI: 1%, 4%; p < 0.001), and further increases were found at RP models (17 ± 3%, 95% CI: 14%, 21%; p < 0.001), but differences between PD and PP models were not significant (p = 0.141). Consequently, there were better improvements in EVR at RD (29 ± 2%, 95% CI: 26%, 31%) and RP (31 ± 6%, 95% CI: 25%, 37%) models than that at PD (10 ± 2%, 95% CI: 8%, 13%) and PP models (7 ± 2%, 95% CI: 4%, 10%; p < 0.001), although no differences between PD and PP models (p = 0.817) and between RD and RP models were detected (p = 1).
Compared with nonassist, the QLM (ml/min) significantly increased at RD (5 ± 2%, 95% CI: 3%, 8%; p = 0.003) and RP model (12 ± 2%, 95% CI: 10%, 14%; p < 0.001), but significantly decreased at PD (−8 ± 2%, 95% CI: −10%, −6%; p < 0.001) and PP model (−7 ± 2%, 95% CI: −9%, −6%; p < 0.001). There was a further increase at RP model than at RD model (p < 0.001), but no differences between PD and PP models (p = 1). There were similar changes in QCA.
Related to nonassist, thermodilution revealed significantly increased COs at RD model (15 ± 3%, 95% CI: 12%, 18%; p < 0.001) and RP models (21 ± 2%, 95% CI: 18%, 23%; p < 0.001), but significantly decreased ones at PD model (−10 ± 3%, 95% CI: −13%, −7%; p < 0.001) and PP model (−13 ± 3%, 95% CI: −16%, −10%; p < 0.001). However, contrary to CO, the LVVO2 showed significant decreases at RD (−7 ± 4%, 95% CI: −11%, −3%; p = 0.005) and RP models (−7 ± 3%, 95% CI: −10%, −3%; p = 0.006), but significant increases at PD model (12 ± 4%, 95% CI: 8%, 18%; p = 0.002) and PP models (22 ± 5%, 95% CI: 17%, 27%; p < 0.001). Those resulted in enhanced CMEs at RD (21 ± 7%, 95% CI: 14%, 28%; p = 0.001) and RP models (25 ± 5%, 95% CI: 20%, 31%; p < 0.001) and damaged ones at PD (−14 ± 4%, 95% CI: −10%, −19%; p < 0.001) and PP models (−21 ± 5%, 95% CI: −16%, −26%; p < 0.001), although there were no differences between PD and PP models (p = 0.225) and between RD and RP models (p = 1). Cardiac output/LAP, another LV functional parameter, also showed similar changes.
As shown in Figure 4, CME was significantly correlated with CO/LAP (r = 0.828, p < 0.001), CO (r = 0.915, p < 0.001), ELVW (r = 0.960, p < 0.001), QCA (r = 0.820, p < 0.001), and LVVO2 (r = −0.940, p < 0.001), rather than QLM (r = 0.692, p < 0.001), MADP (r = 0.421, p = 0.021), or EVR (r = 0.483, p = 0.016). Although it was strongly correlated with CME versus whether or not there was carotid artery reflux (r = 0.873, p < 0.001), there was significantly attenuated relationship between CME and amount of reflux (r = −0.581, p = 0.003). Accidentally, well correlation of QLM and MADP failed to be found (r = 0.428, p = 0.02).
Counterpulsation provides many important hemodynamic benefits, including LV unloading and improvement of myocardial and end-organ perfusion. Although IABP has become the most widely available counterpulsation device, PACD is another promising one, as summarized in Table 2.
In a computer model of the human cardiovascular system, Giridharan et al. tested PACD assistance with two timing protocols as follows: 1) filling before LV systole and ejection after aortic valve closure (at the end of isovolumetric relaxation); 2) filling at beginning of LV systole and ejection just at aortic valve closure. And it was found that the first one was optimal for both LV unloading and coronary arterial perfusion.11 In the present study, four timing algorithms were investigated to get a more comprehensive and accurate result. In an animal model of HF, our data revealed that the optimal assistant algorithm of PACD was filling at beginning of LV systole and ejection at the end of isovolumic relaxation phase, which was different from results reported by Giridharan et al. in a computer mock circulation, perhaps due to difference in the anastomosis sites. In addition, if suboptimal assistant algorithm was performed, the assistant benefits of PACD would be attenuated or even reversed with a further impairment of LV function, which indicated that the timing algorithm was rather important for hemodynamic effects.
We speculate that the following mechanisms may account for these results. As shown in Figure 2, PACD filling before LV systole leads to reflux of blood from distal part of the aorta (at PD and PP models), which cannot stop immediately just after filling finishes due to inertia. After LV systole begins, crash between the reflux and forward blood flow will increase LV afterload. Therefore, the end of PACD filling immediately followed by LV systole may cause the most serious injury of cardiac function. Synchronization of PACD filling and LV systole avoids the reflux and crash, consequently unloads LV (at RD and RP models). PACD ejection after isovolumetric relaxation provides more effective coronary perfusion than ejection at the dicrotic notch, because of lower resistance to the coronary and endocardial perfusion. Therefore, the RP model unloaded LV, decreased myocardial oxygen demands, and increased CO and coronary perfusion most effectively in this experiment.
Optimal aortic pressure waveforms with IABP assistance are described as follows22: the peak diastolic augmentation should be greater than the unassisted systolic pressure; the assisted aortic end-diastolic pressure should be less than the unassisted values. In the present study of PACD support, however, similar aortic pressure waveforms found at PP and PD models failed to provide optimal circulatory assistant effects, even increased myocardial oxygen demands, and worsen LV systolic function. On the contrary, the waveform without a steep drop lower than baseline pressure at the end of diastolic phase, detected at RP and RD models, effectively unloaded LV and promoted cardiac function. In addition, our results revealed weak relations between aortic pressure-related parameters (e.g., EVR and MDAP) and cardiac function-related parameters (e.g., CME, LVVO2, and CO). Therefore, it is strongly suggested that aortic pressure waveforms and its related parameters cannot provide correct guidance for PACD to adjust the timing of balloon inflation or deflation. There has been no similar report. In previous animal and clinical studies of PACD, these parameters were used as important indexes, it is thus speculated that PACD might fail to provide optimal assistance benefits due to suboptimal triggering algorithm, which may be one of reasons why PACD has not been used widely.
Despite encouraging results, there are still remaining questions and considerations that need to be further addressed. These include how to determine the optimal stroke volume of PACD and to find out the scope for application of PACD. Finally, the present study was performed in animal models. The differences between the human circulatory system and sheep one should be taken into account.
In conclusion, we have demonstrated in a sheep model of acute ischemic HF that the optimal triggering algorithm of PACD was filling at beginning of LV systolic phase and ejection at the end of isovolumic relaxation phase. Filling before LV systolic phase increased LV afterload and aggravated HF. Carotid artery flow and ventricular pressure rather than aortic pressure waveforms and its related parameters may be more competent to adjust the timing of balloon inflation or deflation. The present study provided guidance for clinical application of PACD.
The authors gratefully acknowledge the technical assistance of cardiac surgeons (Zhuo Li, Chun-Mao Lu, and Bo Liu).
1. Deng MC. Orthotopic heart transplantation: Highlights and limitations. Surg Clin North Am. 2004;84:243–255
2. Stevenson LW, Rose EA. Left ventricular assist devices: Bridges to transplantation, recovery, and destination for whom? Circulation. 2003;108:3059–3063
3. Baughman KL, Jarcho JA. Bridge to life—Cardiac mechanical support. N Engl J Med. 2007;357:846–849
4. Nanas JN, Nanas SN, Charitos CE, et al. Hemodynamic effects of a counterpulsation device implanted on the ascending aorta in severe cardiogenic shock. ASAIO Trans. 1988;34:229–234
5. Nanas JN, Mason JW, Riebman JB, Charitos CE, Burns GL, Olsen DB. Preclinical evaluation of the abdominal aortic counterpulsation device. Am Heart J. 1988;116:1003–1008
6. Nanas JN, Lolas CT, Charitos CE, et al. A valveless high stroke volume counterpulsation device restores hemodynamics in patients with congestive heart failure and intractable cardiogenic shock awaiting heart transplantation. J Thorac Cardiovasc Surg. 1996;111:55–61
7. Charitos CE, Nanas JN, Kontoyiannis DA, et al. The efficacy of the high volume counterpulsation technique at very low levels of aortic pressure. J Cardiovasc Surg (Torino). 1998;39:625–632
8. Terrovitis J, Charitos Ch, Dolou P, et al. The paraaortic counterpulsation device implanted on the ascending aorta is superior to an equal volume intraaortic balloon pump: Experimental study [abstract]. J Am Coll Cardiol. 2002;39:199A
9. Nose Y, Schamaun M, Kantrowitz A. Experimental use of an electronically controlled prosthesis as an auxiliary left ventricle. Trans Am Soc Artif Intern Organs. 1963;9:269–274
10. Drakos SG, Terrovitis JV, Anastasiou-Nana MI, Nanas JN. Reverse remodeling during long-term mechanical unloading of the left ventricle. J Mol Cell Cardiol. 2007;43:231–242
11. Giridharan GA, Pantalos GM, Litwak KN, Spence PA, Koenig SC. Predicted hemodynamic benefits of counterpulsation therapy using a superficial surgical approach. ASAIO J. 2006;52:39–46
12. Terrovitis JV, Charitos CE, Tsolakis EJ, et al. Superior performance of a paraaortic counterpulsation device compared to the intraaortic balloon pump. World J Surg. 2003;27:1311–1316
13. Ramnarine IR, Capoccia M, Ashley Z, et al. Counterpulsation from the skeletal muscle ventricle and the intraaortic balloon pump in the normal and failing circulations. Circulation. 2006;114(1 Suppl):I10–I15
14. Guyton AC, Hall JE Textbook of medical physiology. 19969th ed Philadelphia, PA WB Saunders
15. Reesink KD, Dekker AL, Van Ommen V, et al. Miniature intracardiac assist device provides more effective cardiac unloading and circulatory support during severe left heart failure than intraaortic balloon pumping. Chest. 2004;126:896–902
16. Goldstein AH, Monreal G, Kambara A, et al. Partial support with a centrifugal left ventricular assist device reduces myocardial oxygen consumption in chronic, ischemic heart failure. J Card Fail. 2005;11:142–151
17. Goldstein AH, Pacella JJ, Clark RE. Predictable reduction in left ventricular stroke work and oxygen utilization with an implantable centrifugal pump. Ann Thorac Surg. 1994;58:1018–1024
18. Pleger ST, Most P, Boucher M, et al. Stable myocardial-specific AAV6-S100A1 gene therapy results in chronic functional heart failure rescue. Circulation. 2007;115:2506–2515
19. Collins P, Mosca L, Geiger MJ, et al. Effects of the selective estrogen receptor modulator raloxifene on coronary outcomes in the Raloxifene Use for The Heart trial: Results of subgroup analyses by age and other factors. Circulation. 2009;119:922–930
20. Chandra D, Kar B, Idelchik G, et al. Usefulness of percutaneous left ventricular assist device as a bridge to recovery from myocarditis. Am J Cardiol. 2007;99:1755––1756
21. Trabattoni P, Zoli S, Dainese L, Spirito R, Biglioli P, Agrifoglio M. Aortic dissection complicating intraaortic balloon pumping: Percutaneous management of delayed spinal cord ischemia. Ann Thorac Surg. 2009;88:e60–e62
22. Trost JC, Hillis LD. Intra-aortic balloon counterpulsation. Am J Cardiol. 2006;97:1391–1398