Mechanical Circulatory Support of a Univentricular Fontan Circulation with a Continuous Axial-Flow Pump in a Piglet Model

Abstract

Despite the significant contribution of the Fontan procedure to the therapy of complex congenital heart diseases, many patients progress to failure of their Fontan circulation. The use of ventricular assist devices to provide circulatory support to these patients remains challenging. In the current study, a continuous axial-flow pump was used to support a univentricular Fontan circulation. A modified Fontan circulation (atrio-pulmonary connection) was constructed in six Yorkshire piglets (8–14 kg). A Dacron conduit (12 mm) with two branches was constructed to serve as a complete atrio-pulmonary connection without the use of cardiopulmonary bypass. The Impella pump was inserted into the conduit through an additional Polytetrafluoroethylene (PTFE) graft in five animals. Hemodynamic data were collected for 6 hours under the supported Fontan circulation. The control animal died after initiating the Fontan circulation independent of resuscitation. Four pump supported animals remained hemodynamically stable for 6 hours with pump speeds between 18,000 rpm and 22,000 rpm (P1–P3). Oxygen saturation was maintained between 95% and 100%. Normal organ perfusion was illustrated by blood gas analysis and biochemical assays. A continuous axial-flow pump can be used for temporal circulatory support to the failing Fontan circulation as “bridge” to heart transplantation or recovery.

The Fontan procedure, or Fontan/Kreutzer procedure, is a palliative surgical procedure used in children with complex congenital heart defects. It involves diverting the venous blood from the right atrium to the pulmonary arteries without passing through the morphologic right ventricle (RV). It was initially described in 1971 by Fontan and Kreutzer separately as a surgical treatment for tricuspid atresia.¹ In addition to the original atrio-pulmonary connection (RA-PA connection) described by Fontan and Kreutzer, several revisions, such as the bidirection Glenn (hemi-Fontan) and total cavopulmonary connection, were further developed to palliate patients with univentricular physiology.² The RA-PA connections have the disadvantage of consuming more energy, providing less effective blood flow to the pulmonary artery and subsequent arrhythmia. Current practice prefers the modified cavopulmonary connections to the original atrio-pulmonary Fontan connection.^3–5

A significant number of palliated patients have early failure and late hemodynamic complications, including ventricular failure, atrioventricular valve regurgitation, atrial arrhythmias, pleural and pericardial effusions, and protein-losing enteropathy.^{6, 7} Medical treatment options are limited in these patients.⁸ Surgical options include revision of the Fontan pathway in some selected cases or cardiac transplantation, which carries increased risk in this patient population.^{7, 9}

Mechanical circulatory support for left ventricular failure with pulsatile and, more recently, axial flow pumps has become quite common clinically.^{10, 11} However, the use of mechanical circulatory support for patients with the Fontan circulation remains challenging.¹² In the current study, a continuous axial-flow pump (Impella LD; Abiomed, Danvers, Massachusetts) was used to support a model of failing Fontan circulation in a piglet animal model. The hemodynamics of the Fontan circulation with the pump support and recirculation in the conduit were studied to evaluate the feasibility and efficiency of this new application of a miniature axial flow pump in the Fontan circulation.

Methods

Surgical Procedure

Six Yorkshire piglets (8.8 ± 0.9 kg) bred for laboratory research (Thomas Morris, Reisterstown, MD) were used in this study. All the surgical procedures and animal care were carried out according to the approved protocol by the Institutional Animal Care and Use Committee of the University of Maryland School of Medicine. During the course of the animal experiments, all animals received humane care in accordance with the “Guide for Care and Use of Laboratory Animals” (NIH publication 86-23, revised 1996). Anesthesia was induced with fentanyl (10 μg/kg) and midazolam (0.2 mg/kg). Ventilation was started after intubation with 4F or 5F endotracheal tubes. Isoflurane (1–3%) was administered during ventilation. Electrocardiography, pulse oximetry, and body temperature of the animals were monitored during the surgery. The intravenous (IV) and arterial lines were established through the femoral vein and artery by a cut-down procedure.

A median sternotomy was made and the pericardium was opened to expose the heart. Systematic anticoagulation was achieved by administration of heparin (50–200 unit/kg) to produce the accelerated coagulation time (ACT) of 200 seconds or greater. A Dacron conduit with two branches was constructed to serve as a complete atrio-pulmonary connection without the use of cardiopulmonary bypass. The distal end of the conduit was anastomosed to the main PA in the end-to-side fashion. The proximal branches were anastomosed end-to-side to the superior vena cava (SVC) and the inferior vena cava (IVC). Two cavo tapes were placed around the SVC and IVC between the atrium and the SVC/IVC anastomoses to stop the flow from the caval veins to the right atria. In the pump supported animals, a micro-pump was inserted into the conduit through an additional 8 mm PTFE graft placed on the conduit. After the two cavo tapes were tied, the total venous blood was diverted from the SVC/IVC to PA circulation through the Y-shaped graft. The pump impeller was then initialed at 20,000 rpm 5 minutes after the Fontan circulation started. The photograph and the schematic diagram of the supported Fontan circulation were shown inFigure 1. A 14 mm flow probe (Transonic Systems Inc., Ithaca, NY) was placed around the graft to measure the device flow rate. Hemodynamics and blood gas were monitored for 6 hours after the initiation of mechanical circulatory support. During the circulatory support, the pump speed was set at 18,000–22,000 rpm (P1–P3) to achieve the highest pump flow without IVC collapse. Heparin was administrated continuously to maintain the ACT at 150–200 seconds during the study period.

Figure 1: The construction of the Fontan circulation and the implantation of the Impella pump. Left, the photo of the operation. Right, the schematic diagram of the supported Fontan circulation. SVC, superior vena cava; IVC, inferior vena cava; PA, pulmonary artery.

The central venous pressure (CVP) and arterial blood pressure (ABP) were measured through the IV line and arterial line which were connected to the pressure transducers (Edwards Lifescience, Irvine, California). Two 16-G IV catheters (Surflash, Terumo Medical Corporation, NJ) were directly punctured into the RV to measure the right ventricle pressure or the pulmonary artery pressure (PAP) after passage through the pulmonary valve. All the pressures were continuously monitored with a patient monitor (Spacelabs Healthcare, Snoqualmie, WA). The pump flow was recorded from the device controller. Concerning on the recirculation in the conduit around the pump, the actual pump flow generated by the pump was monitored using a 14 mm ultrasonic flow probe (Transonic Systems Inc.) placed on the distal graft. Hemodynamic variables were recorded using a Dataq data acquisition system (Dataq Instruments, Inc., Akron, OH) and analyzed using a custom-made matlab program (MathWorks, Natick, Massachusetts). The hemodynamic data were analyzed at baseline, 5 minutes after initialing of the Fontan circulation, and every 2 hours after mechanical support.

Blood samples from the arterial line and PA were collected at baseline and every hour after initiation of the Fontan circulation. The blood gas analysis was performed every hour using a blood gas analyzer (Stat Profile Phox Plus L; Nova Biomedical, Waltham, Massachusetts) to evaluate the adequate oxygenation during the Fontan circulation.

Statistical Analysis

Data are presented as mean ± standard error (SE) of the mean (mean ± SE). A value of p < 0.05 was considered statistically significant. Differences between the different time points were evaluated using the Student’s t-test.

Results

Study Summary

The control animal experienced cardiopulmonary failure 30 minutes after initialing the Fontan circulation and was euthanized. The micro-pump was successfully implanted in five animals after construction of the Fontan circulation. One was euthanized 2 hours after mechanical support because of bleeding from the conduit. The other four piglets with the Fontan circulation were mechanically supported for 6 hours (Table 1).

Table 1: Overview of All Animals with the Fontan Circulation

Hemodynamics

For the control animal, the circulation failed quickly in 30 minutes with the evidence of an elevated CVP (more than 30 mm Hg) and decreased systemic blood pressures (lower than 30 mm Hg, data not shown). In the pump support group, the micro-pump was initiated 5 minutes after the Fontan circulation was established. In this group, the systolic ABP dropped from 82.0 ± 8.6 mm Hg to 43.8 ± 13.4 mm Hg and the CVP increased from 6.0 ± 1.6 mm Hg to 13.3 ± 3.8 mm Hg (p < 0.05) in 5 minutes after initialing the Fontan circulation. The heart rate dropped from 124 ± 13 beats/min to 97 ± 10 beats/min, although there were no significant changes compared to the baseline. The hemodynamics of the animals became stabilized immediately when the mechanical circulatory support was initialized. After 2 hours, the hemodynamics were similar to those at the baseline. The hemodynamics remained stable until the animals were euthanized electively (Table 2). The typical tracing of the hemodynamic variables in these animals is shown in Figure 2.

Table 2: Hemodynamic Variables

Pump Performance for Mechanical Circulatory Support

The pump generated flow and the actual conduit forward flow to the PA (as measured by the flow probe from the distal of the conduit) were monitored continuously for 6 hours. The pump flow increased from 1.19 ± 0.09 L/min to 3.05 ± 0.05 L/min at increasing speeds from 18 krpm to 24 krpm (P1–P4) (Figure 3A). The actual conduit forward flow also increased from 0.80 ± 0.03 L/min to 1.44 ± 0.01 L/min (Figure 3B). The recirculation flow was calculated by subtracting the actual conduit forward flow from the pump generated flow. We found that the recirculation flow was also increased with the increase in the pump speed. Without a valve, nearly half blood flow was recirculated back in the space between the conduit and the outer housing of the Impella pump (Figure 3C).

Figure 2: The typical tracing of the hemodynamic variables in the supported-Fontan animals. CVP, central venous pressure; ABP, arterial blood pressure; PAP, pulmonary artery pressure; RVP, right ventricle pressure.

Figure 3: Pump performance. A: The pump flow measured by the device sensor. B: The actual conduit forward flow as measured by the flow probe from the distal of the conduit. C: The recirculation flow.

Figure 4: Metabolic data of the supported-Fontan animals. A: Arterial O₂ saturation. B: Et CO₂. C: Arterial lactate.

Metabolism

Arterial pH remained within the normal range during the pump support study period. Arterial O₂ saturation dropped to 82 ± 6 at 5 minutes after the Fontan circulation was established and elevated to 95–100% with the pump support (Figure 4A). Mixed SaO₂ had no significant change during the pump support study period. Et CO₂ decreased from 37 ± 3% at baseline to 28 ± 2% initially after the Fontan circulation was established and increased to normal with the pump support (Figure 4B). Arterial lactate values increased after initialing of the Fontan circulation, but no significant difference was observed during the pump support study period (Figure 4C).

Discussion

Over the past 30 years, the Fontan procedure, its modifications, and extension to cavopulmonary connections have offered a life-saving treatment for devastating anomalies. However, the limitations of Fontan circulation and the need for circulatory support are evident.¹² Unfortunately, the application of mechanical circulatory support to single ventricle physiology has been limited by a lack of suitable devices. In an effort to broaden the use of existing devices for the potential support of Fontan circulations, this study used a piglet model of Fontan circulation with RA-PA connection.

It remains a challenging task to create an animal model of true sustained failing Fontan circulation. A chronic true failing Fontan experimental model has never been reported.¹³This is anticipated because of the technical difficulties in establishing Fontan circulation in a pre-existing normal heart. Therefore, the studies of Fontan circulation has substantially relied upon acute experimental studies in total right heart bypass operations, which can be done by direct tricuspid valve occlusion^{14, 15}or discontinuing the inflow of the right atrial with cavo tapes placed in the proximal IVC and SVC.^{16, 17}In our study, we used the latter technique to exclude blood flow to the RV. A Y-shaped conduit was used to connect the caval veins to the pulmonary artery as others described.^{14, 16} Advantages of this model include the avoidance of cardiopulmonary bypass, which would lead to significant volume shifts and transfusion requirements, likely obscuring results. In addition, the micro-pump is easily inserted in either vascular structures or grafts, eliminating the need for additional cannulation and decreasing surgery time.

In a normal heart, the mean CVP is less than 10 mm Hg and the mean PAP is at least 15 mm Hg. The paradox of the Fontan circulation is that it imposes caval hypertension as well as relative pulmonary arterial hypotension. A mechanical device capable of producing a decrease of 5 mm Hg in inferior caval pressure and an increase of 5 mm Hg in the pulmonary arteries could potentially reverse the Fontan paradox.⁵ Many studies have used ventricular assist devices (VADs) in models of complete right heart bypass and showed hemodynamic and pulmonary support.^{18, 19} Rodefeld et al. showed that two axial flow Hemopumps (no longer commercially available) within both the SVC and IVC maintained hemodynamic stability and oxygenation while decreasing arterial lactate levels in an ovine bicaval-PA model.²⁰They suggested that a small diameter, percutaneous axial flow pump positioned in a relatively much larger diameter central vein requires a barrier to recirculation to ensure effective flow. However, barriers to recirculation are problematic with respect to thrombus formation, especially in the low pressure Fontan venous pathway. Our results indicate the ability of the Impella pump to augment Fontan circulation with stabilization of hemodynamics to baseline condition within 2 hours. This effect was sustained for 6 hours and included normalization of oxygenation, end-tidal CO₂, and arterial lactate. Although nearly 50% flow was recirculated around the pump, the axial flow pump is capable to provide enough forward power to improve the Fontan circulation. The hemodynamic and biologic results of this study suggest a novel approach to improving single-ventricle physiology. A barrier to recirculation may not be needed according to our study. In a mock circulatory system of the Fontan circuit, three types of Impella pumps (Impella 2.5, Impella 5.0, and Impella RP) were evaluated as cavopulmonary assistance for a failing Fontan.²¹ The right-ventricular Impella device could provide improved performance by directing flow into the pulmonary artery, resulting in modest decreases in CVP. Supported by this precedent, our work continues to show not only that mechanical circulatory support can support a failing Fontan circulation short term but also that an endovascular pump maybe suitable for restoration of physiologic venous pressure, oxygenation, and systemic perfusion in permitting Fontan anatomies. A suggested clinical application is shown in Figure 5 if a percutaneous pump-forward device, such as the Impella RP, is used. This device functioned not as a VAD but as a cavopulmonary assist device (CPAD), actively pumping venous return to the PA and serving to decrease CVP and increase pulmonary flow. This approach has the potential to alleviate the profound venous congestion, hypoxemia, and heart failure inevitable without transplant. While extrapolative, this decrease in CVP and increase in PA flow could not only reduce hepatic congestion and protein-losing enteropathy, but also improve gas exchange. In addition, the increased pulmonary perfusion would supplement left heart preload, enhancing cardiac output and systemic perfusion.

Figure 5: Possible clinical application of a percutaneous catheter-based ventricular assist device (Impella RP) in patients with the RA-PA Fontan connection.

Other devices have also been tried to support the Fontan circulation. The Heartmate II was used to support a model using grafts to connect both the SVC and the IVC to PA. While only supporting IVC to PA flow, the Heartmate II was able to modulate the decrease in cardiac output and ABP and the IVC hypertension seen with the unsupported model.²² In a sheep model, a study using the dual lumen Avalon cannula for extracorporeal membrane oxygenation (ECMO) and a Centrimag pump was the first to take advantage of a surgically placed conduit as a site of intervention. The cannula was placed within a cavopulmonary graft, connecting both cavae with a bridge graft to the PA. With dual proximal/SVC and distal/IVC inflow and outflow to the PA bridge graft, this approach was able to hemodynamically support a failing Fontan model for 2 hours.²³ Rodefeld et al. has also designed a pediatric viscous impeller pump with a biconical design allowing it both draw and pump blood in two directions using a single pump head.²⁴ Small enough to fit in the cavopulmonary junction, in vitro studies suggest this pump can support flows up to 4 L/min with minimal shear stress. Its effectiveness in vivo remains to be seen.

Limitations of this study include an incomplete model of univentricular physiology. While the right heart is completely excluded, our model does not illustrate the chronic effects of Fontan circulation, mainly increased pulmonary vascular resistance and systemic venous hypertension. Further work will examine the micro-pump support of a Fontan circulation in a long-term animal model.

Conclusion

A miniature axial-flow pump was demonstrated to be capable of providing the hemodynamic support and maintaining normal oxygenation and systemic perfusion of a failing Fontan circulation. This approach may be used to sustain more single ventricle patients to transplantation or to recovery.