Secondary Logo

Journal Logo

Adult Circulatory Support

Distinct Effects of Left or Right Atrial Cannulation on Left Ventricular Hemodynamics in a Swine Model of Acute Myocardial Injury

Esposito, Michele L.; Shah, Nimish; Dow, Sam; Kang, Sandra; Paruchuri, Vikram; Karas, Richard H.; Kapur, Navin K.

Author Information
doi: 10.1097/MAT.0000000000000416


Use of acute mechanical circulatory support (AMCS) devices to provide hemodynamic support during elective high-risk percutaneous coronary, structural, and electrophysiology procedures is growing exponentially.1–3 In each of these settings, acute impairment of myocardial contractility increases left ventricular (LV) end-diastolic pressure and volume leading to higher LV wall stress and myocardial oxygen demand. Two primary goals of AMCS systems are to 1) increase mean arterial pressure (MAP) and vital organ perfusion (circulatory support) and 2) to reduce ventricular pressure and volume, thereby reducing wall stress, stroke work, and myocardial oxygen consumption (LV unloading). Commonly used AMCS devices include 1) an axial flow catheter in the left ventricle (Impella; Abiomed Inc, Danvers, MA), 2) a left atrial to femoral artery (LA–FA) extracorporeal centrifugal pump (LA–FA bypass; TandemHeart; Cardiac Assist Inc, Pittsburgh, PA), or 3) right atrial to femoral artery (RA–FA) extracorporeal centrifugal pump with an oxygenator (RA–FA bypass; veno-arterial extracorporeal membrane oxygenation; VA-ECMO).4–6

All three of these AMCS systems generate continuous blood flow and reduce native LV work by transferring kinetic energy from a circulating impeller to the blood stream. In contrast to Impella axial-flow catheters, both the TandemHeart and VA-ECMO employ inflow and outflow cannulas connected to an extracorporeal centrifugal pump that drain blood from the heart into the arterial system. A major difference between the TandemHeart and VA-ECMO systems for LV support is the location of the inflow cannula, which is placed in the LA via trans-septal puncture for the TandemHeart system and across the right atrial (RA) for VA-ECMO. While mathematical models have attempted to simulate the hemodynamic effects of these devices,4 an unresolved issue in the field of AMCS is how location of the inflow cannula in the RA or LA impacts LV hemodynamics. In this study, we explore the central hypothesis that inflow cannula location in the LA provides a greater LV unloading compared with RA cannulation in the setting of acute left heart injury.


Studies were conducted in adult, male swine weighing 50–65 kg. All procedures were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and were approved by Tufts Medical Center and the Institutional Animal Care and Use Committee.

Experimental Protocol of Acute Left Ventricular Injury and Mechanical Circulatory Support

Animals were premedicated with intramuscular ketamine (4.5 mg/kg) and xylazine (0.3 mg/kg). General anesthesia was induced and maintained with isoflurane (1–3%). All animals were intubated and mechanically ventilated (Harvard Apparatus Inc, Holliston, MA) to control oxygen and isoflurane gas intake, oxygen saturation, and expired CO2 levels. Surface electrocardiography leads, an orogastric tube, peripheral 18 G venous catheters, pulse oximeter, and a rectal thermometer were placed in all animals. Heating pads were used as needed to maintain a core body temperature >99° F. Vascular access sheaths were then deployed into the right internal jugular vein (10 Fr) and left carotid artery (7 Fr). A 5 Fr-conductance catheter (CD Leycom) was placed into the LV via the left carotid artery. Unfractionated heparin boluses with a goal activated clotting time of 300–400 seconds, a continuous lidocaine infusion, and noradrenaline were initiated in all animals.

In the TandemHeart group (LA–FA bypass), a 21 Fr trans-septal inflow cannula (TandemHeart; Cardiac Assist Inc) was introduced into the LA via the right femoral vein using a trans-septal sheath and dilator (SL-1; Medtronic Inc, Minneapolis, MN). In the VA-ECMO group (RA–FA bypass), a 21 Fr inflow cannula introduced into the RA via the right femoral vein. In both groups, a 17 Fr outflow cannula was deployed into the right femoral artery. The inflow and outflow cannulas were connected to a centrifugal flow pump (TandemHeart; Cardiac Assist Inc, Pittsburgh, PA) and maintained at 3,500 RPMs (Figure 1).

Figure 1.
Figure 1.:
Experimental design. A: A line graph illustrates the timing of LA–FA or RA–FA bypass activation, the timing of LCx occlusion and reperfusion, the ramp protocol, and timing of PVL measurements. Illustrations of (B) RA–FA bypass or VA-ECMO and (C) LA–FA bypass or TandemHeart positioning. Conductance catheters were introduced into the LV via the carotid artery. LA–FA, left atrial to femoral artery; RA–FA, right atrial to femoral artery; LCx, left circumflex; PVL, pressure-volume loop; VA-ECMO, veno-arterial extracorporeal membrane oxygenation; LV, left ventricular.

To explore the impact of device activation in a clinically relevant model of acute LV injury, a 6 Fr Judkins right coronary catheter (Boston Scientific, Marlborough, MA) engaged the left coronary artery via the right femoral artery and baseline angiograms were recorded, while the circulatory support pump was maintained at 3,500 RPMs, which is the lowest flow setting for the centrifugal pump. A 180 cm guidewire was delivered into the distal left circumflex (LCx) artery and a 3.0 × 8 mm coronary balloon inflated in the proximal LCx. The LCx was selected because LV injury can reliably be reproduced with proximal LCx occlusion, mortality is less with acute LCx occlusion than with left anterior descending (LAD) occlusion, and LAD occlusion (anterior MI) often causes right ventricular injury, which could impact our analysis of the LA–FA and RA–FA bypass circuits.7,8 After 30 minutes of LCx occlusion, the LCx was reperfused for an additional 30 minutes, while maintaining both the centrifugal pump at 3,500 RPMs. Hemodynamics were recorded after 30 minutes of reperfusion beginning at the 3,500 RPM setting then after increasing the RPM setting from 3,500 to 5,500, then 7,500 for 10 minutes at each step of the ramp protocol (Figure 1). Thirty minutes after LCx occlusion, the LCx was reperfused and RPMs through the extracorporeal circuit were increased from 3,500 to 5,500, then 7,500 for 10 minutes at each step of the ramp protocol (Figure 1). Device flow was measured using a flow probe attached to the outflow cannula. Hemodynamic data using the conductance catheter was acquired at the end of each ramp step as described below. In the LA–FA bypass group, the LA cannula was clamped after completion of the ramp protocol for 10 minutes, reactivated in the LA for 10 minutes, then retracted into the RA. At the end of the protocol, the circuit was clamped and animals euthanized with pentobarbital and phenytoin. Eight out of 10 animals survived to protocol completion (n = 4/group); two animals died because of ventricular fibrillation during LCx occlusion.

Conductance Catheter Assessment of Left Ventricular Pressure and Volume

Changes in LV pressure and volume were assessed using a 5 Fr-conductance catheter system (Sigma-M; CD Leycom, The Netherlands) deployed via the left carotid into the LV (Figure 1C). Ventricular pressure and volume were measured using a solid-state pressure transducer and dual-field excitation mode, respectively, as previously described.9–11 Parallel conductance was assessed by injecting 20 ml of hypertonic (6%) saline into the right internal jugular vein. Absolute LV volumes were measured by subtracting parallel conductance from total conductance volumes. Native LV stroke volume is calculated as the difference in conductance volumes at dP/dtmax and dP/dtmin. Left ventricular stroke work (LVSW) was calculated as the product of peak LV peak systolic pressure and stroke volume. Preload recruitable stroke work (PRSW) was calculated as the ratio LVSW and LV end-diastolic volume (LVEDV) using the single-beat method as previously reported.12 Arterial elastance (Ea) was calculated as LV end-systolic pressure (LVESP) divided by stroke volume.13,14 Because of the use of large bore cannulas in the inferior vena cava, we were unable to directly quantify pressure–volume areas as a surrogate measure of myocardial oxygen consumption or LV wall stress. For this reason, estimated LV wall stress was calculated as the product of LVEDV and LVESP based on prior reports.15,16

Statistical Analysis

Results are presented as mean ± standard deviation (SD). All data within groups were analyzed by nonparametric two-way repeated measures analysis of variance (ANOVA) on ranks followed by a Holm–Sidak comparison if warranted. All statistical analyses were performed with SigmaStat version 3.1 (Systat Software, Inc, San Jose, CA). An α-level of p < 0.05 was considered to indicate a significant effect or between-group difference.


In the RA–FA bypass group, VA-ECMO generated 1.1 ± 0.1, 2.5 ± 0.1, and 3.5 ± 0.1 liters/minute of flow at 3,500, 5,500, and 7,500 RPMs. Compared with baseline values, RA unloading at 7,500 RPMs did not reduce LV end-diastolic pressure (LVEDP), LVEDV, stroke volume, or LVSW. Right atrial unloading increased MAP, LVESP, and estimated wall stress, but did not alter Ea (Table 1 and Figure 2; see Video, Supplemental Digital Content 1, In the LA–FA bypass group, the TandemHeart generated 0.9 ± 0.3, 2.3 ± 0.3, and 3.6 ± 0.2 L/minute of flow at 3,500, 5,500, and 7,500 RPMs. Compared with baseline values, LA unloading at 7,500 RPMs reduced LVEDP, LVEDV, stroke volume, and LVSW without significantly changing MAP. Estimated wall stress was unchanged and Ea increased with maximal activation of the TandemHeart device compared with baseline values (Table 1 and Figure 2; see Video, Supplemental Digital Content 2,

Table 1.
Table 1.:
Hemodynamic Effects of LA–FA Versus RA–FA Bypass
Figure 2.
Figure 2.:
Hemodynamic effects of LA–FA versus RA–FA bypass. Line graphs showing the effect of LA–FA versus RA–FA bypass on (A) Ea, (B) estimated LV wall stress, and (C) LV stroke volume. LV, left ventricular; RA–FA, right atrial to femoral artery; LA–FA, left atrial to femoral artery; Ea, arterial elastance.

To further explore the hemodynamic impact of LA versus RA inflow, the LA cannula was clamped after completion of the ramp protocol for 10 minutes, reactivated in the LA for 10 minutes, and then retracted into the RA. Compared with the clamped condition, LA–FA bypass reduced LVSW, stroke volume, PRSW, and LVEDP, while increasing Ea. Wall stress was not changed by activating LA–FA bypass. Compared with LA–FA bypass, RA–FA bypass increased LVSW, stroke volume, PRSW, LVEDP, and calculated wall stress. Arterial elastance was reduced with RA–FA bypass compared with LA–FA bypass (Table 2 and Figure 2).

Table 2.
Table 2.:
Hemodynamic Effects of Converting LA–FA to RA–FA Bypass


This is the first report to directly compare the acute effects of RA–FA bypass (VA-ECMO) versus LA–FA bypass (TandemHeart) using an extracorporeal centrifugal pump on LV hemodynamics in a closed chest model of acute LV injury. Our central finding is that at maximal flow rates, RA and LA positioning of the inflow cannula generate distinct hemodynamic effects in a swine model of acute LV injury. Specifically we report that at maximum flow, RA–FA bypass increases MAP and LV end-systolic pressure, while having minimal impact on LV volumes. In contrast, LA–FA bypass reduces LV end-diastolic pressure and volume, native stroke volume, and LV stroke work. By directly reducing LV preload, LA–FA bypass reduces LV stroke work primarily by reducing native LV stroke volume, whereas RA–FA bypass fails to reduce native LV stroke volume or stroke work. Furthermore, we show that RA–FA bypass increases estimated LV wall stress without impacting Ea, whereas LA–FA bypass increases Ea without affecting estimated LV wall stress. These data suggest that LA–FA bypass can provide both circulatory support and LV unloading, whereas RA–FA bypass provides circulatory support without LV unloading. Based on these findings, LA–FA bypass may be a more effective approach to protect the LV than RA–FA bypass during acute circulatory support. These findings have potentially important implications for clinical decision-making and for the design of future percutaneously-delivered circulatory support devices.

We previously reported that the Impella CP provides myocardial protection by reducing LV pressure, volume, and stroke work in preclinical models of acute left heart injury and in acute myocardial infarction.17,18 We now studied the effect of the LA–FA and RA–FA bypass on LV hemodynamics. The TandemHeart (LA–FA bypass) and VA-ECMO (RA–FA bypass) systems share several features. Both employ an extracorporeal centrifugal pump attached to inflow and outflow cannulas that deliver blood into the arterial system. By displacing blood volume from the LA or RA into arterial system, both systems have the potential to increase MAP and LV afterload; however, LV afterload is best defined as stress generated in the wall of the LV during systolic ejection and incorporates vascular resistance, impedance, and elastance.19 Based on the Law of Laplace, afterload can be measured as LV wall stress, which equals LV pressure times the radius during systolic ejection divided by two times the LV wall thickness. Acute changes in LV wall stress (or afterload) can be estimated in the pressure–volume domain as the product of LV end-systolic pressure and end-diastolic volume (Figure 3).15 The term “arterial elastance” (Ea) was first proposed by Sunagawa et al.13,14 and is defined as LV end-systolic pressure divided by stroke volume. Arterial elastance is a component of afterload, but does not always directly correlate with LV afterload or wall stress. Based on our findings, RA–FA and LA–FA bypass have distinct effects on estimated LV wall stress and Ea.

Figure 3.
Figure 3.:
Distinct effects of LA–FA versus RA–FA bypass on Ea and estimated LV wall stress. A: Pressure volume loops demonstrate that maximal activation of RA–FA bypass increases LV-ESP (blue dots) without affecting LV-EDV (green dots). B: Illustrated PV loops show that compared with baseline conditions (black lines), RA–FA bypass (red lines) increases Ea (hashed line), estimated LV wall stress (area within the On versus Off squares), without significantly decreasing LV SV (width of the loop). CD: Maximal activation of LA–FA bypass increases LV-ESP and decreases LV-EDV and SV, thereby increasing Ea without significantly altering LV wall stress. LA, left atrial; RA–FA, right atrial to femoral artery; Ea, arterial elastance; LV, left ventricular; LV-ESP, LV end-systolic pressure; LV-EDV, LV end-diastolic volume; PV, pressure volume; SV, stroke volume; LA–FA, left atrial to femoral artery.

As RA–FA bypass displaces blood from a large venous reservoir into the arterial system without directly unloading the LV, we observed an increase in LV end-systolic pressure without any significant change in LV end-diastolic volume or stroke volume. As a result, RA–FA bypass increased estimated LV wall stress because of a higher LVESP, but Ea was largely unchanged because LV stroke volume was not affected (Figure 3). In contrast, as LA–FA bypass displaces blood from the LA into the arterial system, LV preload is directly reduced and we observed that LVESP was maintained, but LVEDV and LV stroke volume were reduced. As a result, LA–FA bypass did not affect estimated LV wall stress, but increased Ea largely by reducing LV stroke volume (Figure 3). These findings were confirmed using distinct experimental subjects (n = 4/group) and by retracting the LA cannula into the RA, thereby effectively converting LA–FA to RA–FA bypass within the same study subject at the end of the ramp protocol in the LA–FA bypass group (n = 4; Figure 1 and Table 2).

A primary objective of AMCS devices is to unload the LV by reducing wall stress and hence myocardial oxygen demand. This objective is of particular importance during high risk percutaneous coronary intervention (PCI) because increased LV wall stress may promote LV ischemia, which impairs both LV systolic and diastolic function leading to worsening heart failure and an increased propensity for ventricular arrhythmias. Based on our findings, RA–FA bypass (VA-ECMO) can increase MAP, but does so at the cost of increasing LV wall stress. For this reason, RA–FA bypass may not represent an optimal AMCS strategy for patients undergoing high-risk PCI. These findings are consistent with several reports indicating the need for additional measures to mitigate increased LV wall stress in patients supported with RA–FA bypass. In contrast, LA–FA bypass can be used to maintain MAP (circulatory support) and reduce LV stroke work without increasing LV wall stress (LV unloading).

There exist several limitations with this study. First, we did not study the hemodynamics of these devices in the setting of chronic LV failure or cardiogenic shock. Second, the number of animals studied was small. Finally, we did not directly assess the magnitude of LV injury in each animal. To avoid any potential confounding effects of variable LV injury, we studied the effect of retracting the LA cannula into the RA (Table 2) in the same animal and observed a similar pattern of hemodynamic effects with LA versus RA cannulation in distinct animal subjects. Future studies involving a larger number of animals and longer duration of LV injury are required.

In conclusion, these data show that devices designed to remove blood from the LA can effectively reduce LV stroke work while sustaining systemic perfusion by primarily reducing LV volume, whereas devices that remove blood from the RA do not unload the LV and may increase LV wall stress. The technical aspects of device deployment including the use of trans-septal puncture, the need for large bore vascular access, and an institutions capability to support VA-ECMO will continue to guide clinical decision-making. Findings from this analysis may inform the design of future devices and the development of prospective clinical studies to evaluate the role of AMCS devices in the setting of acute LV injury.


1. Stretch R, Sauer CM, Yuh DD, Bonde PNational trends in the utilization of short-term mechanical circulatory support: Incidence, outcomes, and cost analysis.J Am Coll Cardiol20146414071415
2. Khera R, Cram P, Lu X, et al.Trends in the use of percutaneous ventricular assist devices: Analysis of national inpatient sample data, 2007 through 2012.JAMA Intern Med2015175941950
3. Rihal CS, Naidu SS, Givertz MM, et al.Society for Cardiovascular Angiography and Interventions (SCAI); Heart Failure Society of America (HFSA); Society for Thoracic Surgeons (STS); American Heart Association (AHA); American College of Cardiology (ACC)2015 SCAI/ACC/HFSA/STS clinical expert consensus statement on the use of percutaneous mechanical circulatory support devices in cardiovascular care (endorsed by the American Heart Association, the Cardiological Society of India, and Sociedad Latino Americana de Cardiologia Intervencion; affirmation of value by the Canadian Association of Interventional Cardiology-Association Canadienne de Cardiologie d’intervention).J Card Fail201521499518
4. Burkhoff D, Sayer G, Doshi D, Uriel NHemodynamics of mechanical circulatory support.J Am Coll Cardiol20156626632674
5. Kapur NK, Esposito MHemodynamic support with percutaneous devices in patients with heart failure.Heart Fail Clin201511215230
6. Aghili N, Kang S, Kapur NKThe fundamentals of extra-corporeal membrane oxygenation.Minerva Cardioangiol2015637585
7. Baan J, van der Velde ET, de Bruin HG, et al.Continuous measurement of left ventricular volume in animals and humans by conductance catheter.Circulation198470812823
8. Bodi V, Sanchis J, Mainar L, et al.Right ventricular involvement in anterior myocardial infarction: A translational approach.Cardiovasc Res201087601608
9. Ishikawa K, Aguero J, Tilemann L, et al.Characterizing preclinical models of ischemic heart failure: Differences between LAD and LCx infarctions.Am J Physiol Heart Circ Physiol2014 Nov 15307H1478H1486
10. Kass DA, Yamazaki T, Burkhoff D, Maughan WL, Sagawa KDetermination of left ventricular end-systolic pressure-volume relationships by the conductance (volume) catheter technique.Circulation198673586595
11. Schreuder JJ, van der Veen FH, van der Velde ET, et al.Left ventricular pressure-volume relationships before and after cardiomyoplasty in patients with heart failure.Circulation19979629782986
12. Ferrazzi P, Senni M, Iascone MR, et al.Implantation of an elastic ring at equator of the left ventricle influences cardiac mechanics in experimental acute ventricular dysfunction.J Am Coll Cardiol20075017911798
13. Sunagawa K, Maughan WL, Burkhoff D, Sagawa KLeft ventricular interaction with arterial load studied in isolated canine ventricle.Am J Physiol19832455 Pt 1H773H780
14. Sunagawa K, Maughan WL, Sagawa KOptimal arterial resistance for the maximal stroke work studied in isolated canine left ventricle.Circ Res198556586595
15. Gaasch WH, Battle WE, Oboler AA, Banas JS Jr, Levine HJLeft ventricular stress and compliance in man. With special reference to normalized ventricular function curves.Circulation197245746762
16. Izzi G, Zile MR, Gaasch WHMyocardial oxygen consumption and the left ventricular pressure-volume area in normal and hypertrophic canine hearts.Circulation19918413841392
17. Kapur NK, Paruchuri V, Pham DT, et al.Hemodynamic effects of left atrial or left ventricular cannulation for acute circulatory support in a bovine model of left heart injury.ASAIO J201561301306
18. Kapur NK, Qiao X, Paruchuri V, et al.Mechanical pre-conditioning with acute circulatory support before reperfusion limits infarct size in acute myocardial infarction.JACC Heart Fail20153873882
19. Laurent S, Cockcroft J, Van Bortel L, et al.Expert consensus document on arterial stiffness: Methodological issues and clinical applications.Eur Heart J20062725882605

circulatory support; heart failure; hemodynamics

Copyright © 2016 by the ASAIO