Abdominal Positioning of the Next Generation Intra-Aortic Fluid Entrainment Pump (Aortix) Improves Cardiac Output in a Swine Model of Heart Failure

Background

Acute heart failure refractory to medical therapy is a major cause of morbidity and mortality. The Aortix device (Procyrion Inc) is a percutaneously delivered, entrainment pump positioned in the descending aorta. Using the newest generation Aortix device in eight adult male Yorkshire swine, we tested the hypothesis that positioning in the abdominal aorta may provide superior hemodynamic effects than thoracic positioning in a swine model of post-infarct LV injury. Abdominal activation generated significantly larger trans-aortic gradients (proximal minus distal mean aortic pressures) than thoracic positioning at all pump speeds. Compared to baseline values, activation in the abdominal, not thoracic, position significantly increased cardiac output, reduced arterial elastance, and systemic vascular resistance at low speeds. Compared to baseline values, abdominal activation also increased trans-pulmonary pressure gradients at medium and high speed, which was driven by trends towards higher mean pulmonary artery pressure and lower pulmonary capillary wedge pressure. This is the first report to determine that in contrast to thoracic positioning, abdominal positioning of the newest generation Aortix device reduces LV afterload and increases cardiac output at low speeds. These findings have potentially important implications for the design of early clinical studies by suggesting that device position and speed are major determinants of improved hemodynamic efficacy.

Introduction

Nearly 2 million patients are admitted with the primary diagnosis of acute decompensated heart failure each year in the United States alone ¹. Use of acute mechanical circulatory support (AMCS) pumps has increased, however these devices are indicated for decompensated heart failure without cardiogenic shock ^2–4. For these reasons, novel devices for medically refractory heart failure are needed.

The Aortix device is a percutaneously-delivered, catheter-mounted, axial-flow pump (Aortix, Procyrion Inc) positioned in the descending aorta. The device is a small rotating impeller mounted within a self-expanding nitinol strut anchoring system that is connected to a motor controller by a thin, flexible, electrical wire. When activated, high-velocity jets at its outlet accelerate native aortic flow and entrain blood flow thereby creating a trans-aortic gradient. In 2013, preclinical testing with the first generation Aortix device showed a modest increase in cardiac output when positioned in the thoracic aorta ⁵. Since then, the Aortix device has been modified to include a shortened cannula-impeller unit and altered aortic anchoring tines to minimize shear stress, improve pump performance, and enhance deliverability (Figure 1A).

Figure 1.

A) Illustrations of the first and fourth generation Aortix devices; B) Preclinical study protocol; C) Fluoroscopic image showing the Aortix device; D) Trans-aortic pressure tracings during abdominal activation of the Aortix device

We postulated that positioning the Aortix device in the abdominal aorta may provide superior hemodynamic effects than thoracic positioning. We tested this hypothesis using the newest generation Aortix device in a swine model of post-infarct heart failure.

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Experimental Protocol

Aortix implant studies were conducted in eight adult male Yorkshire swine. Four adult male Yorkshire swine serving as controls for baseline hemodynamic assessment of heart failure. All animals were of the same size, weight, and breed and had similar control of drugs and volume. The study protocol was approved by the Institutional Animal Care and Use Committee (IACUC) at Tufts Medical Center. All experiments were performed according to the committee’s guidelines. Experimental details are reported in Figure 1B. Isoflurane (1-2%), continuous lidocaine infusion (12mg/min), and norepinephrine (0.5-0.8 mcg/min) were initiated in all animals. Briefly, the mid-LAD was occluded after the second diagonal branch using an angioplasty balloon for 120 minutes delivered via the right femoral artery. Following reperfusion, all animals were recovered and monitored for 28 days. After 28 days, animals were anesthetized, ventilated, and instrumented with a left ventricular conductance catheter (Millar Inc), pulmonary artery catheter (Edwards), a Doppler flow wire in the left circumflex artery (Volcano; Phillips), and pressure transducers in the right carotid artery, aortic root, and distal aorta. The Aortix device was advanced into the descending aorta via a 14Fr sheath in the left femoral artery (Figure 1B). Thoracic positioning was defined as above the diaphragm between the T5-T6 vertebral bodies. Abdominal positioning was defined as below the diaphragm between the T10-T-12 vertebral bodies. The Aortix device was activated and hemodynamic measurements were again obtained after 10 minutes at incremental ramp speeds: medium (28-30K), high (34-37K) in thoracic position (n = 5) and low (22-25K), medium, high in the abdominal position (n = 5). Low RPMs were not recorded in the thoracic position. Animals were euthanized at the end of study protocol (Figure 1B-C).

Hemodynamic Assessment

Changes in LV pressure and volume were measured using a solid-state pressure transducer and dual-field excitation mode respectively as previously described ^6–8. Pulmonary artery catheter indices were quantified at each step in the ramp protocol (Figure 1B).

Statistical Analysis

Hemodynamic data from the total 8 animals receiving the Aortix device were grouped based on the position of the device, which yielded 5 hemodynamic datasets for each of the abdominal and thoracic groups. Baseline hemodynamic data were compared between the 4 control animals and abdominal and thoracic groups prior to Aortix implantation. Baseline hemodynamic data were compared between the 4 control animals and abdominal and thoracic groups prior to Aortix implantation. All hemodynamic data were then grouped into four subsets of pump speed: baseline (no pumping), low (22-25K), medium (28-30K), and high (34-37K). For abdominal and thoracic positions, intra-position data was analyzed using a repeated measure one-way ANOVA. For inter-position analysis (abdominal vs thoracic), a student’s T-test was performed comparing relative changes at each pump speed from baseline. Specifically, comparisons included baseline versus each pump speed group (Baseline vs. Low; Baseline vs. Medium; Baseline vs. High) and comparisons between each pump speed group (Low vs. Medium; Low vs. High; Medium vs. High). Thoracic and abdominal position data at baseline and any given pump speed were compared using a t-test. All statistical analyses were performed using GraphPad Prism, Version 7 (Graph-pad Software, Inc). Data are expressed as mean ± standard deviation for continuous variables. A p-value < 0.05 indicated statistical significance.

Results

The Aortix device was successfully implanted in all eight swine subjects across the three study protocols (Figure 1C). Data were analyzed for the abdominal (n=5) or thoracic (n=5) positions. All hemodynamic data are shown in Table 1. When compared to healthy control animals, both abdominal and thoracic animals had reduced LV stroke work (SW) (Control: 4153 ± 729 mmHg·mL; vs. Abdominal: 2295 ± 510 mmHg·mL, p = 0.003; vs. Thoracic: 2904 ± 729 mmHg·mL, p = 0.026) lower aortic systolic pressures (Control: 108 ± 4 mmHg; vs. Abdominal: 76 ± 7 mmHg, p = 0. 0006; vs. Thoracic: 89 ± 10 mmHg, p = 0.019), and a trend towards lower PCWP in control animals (Control: 6 ± 1 mmHg; vs. Abdominal: 9 ± 1 mmHg, p = 0.099; vs. Thoracic: 9 ± 2 mmHg, p = 0.099). The abdominal and thoracic group animals had significantly different AOSP, LVESV, LVEDV at baseline.

Table 1.

Hemodynamic Effects of Aortix Activation

	Control	Thoracic			Abdominal
		Baseline	28-30K	34-37K	Baseline	22-25K	28-30K	34-37K
Heart Rate (bpm)	92 ± 15	65 ± 10	77 ± 23	79 ± 29	67 ± 11	73 ± 12	75 ± 9	83 ± 18
Proximal Ao MAP (mmHg)	89 ± 6	74 ± 13	73 ± 17	70 ± 18	64 ± 8 ‡	59 ± 14	58 ± 12	46 ± 20
Distal Ao MAP (mmHg)	N/A	69 ± 12	79 ± 17	80 ± 17	63 ± 7	71 ± 10	78 ± 5 *	80 ± 10
Trans-Ao Gradient (mmHg)	N/A	−5 ± 2	6 ± 1 *	10 ± 1 *	−1 ± 2	12 ± 6 *†	20 ± 9 *†ˠ	34 ± 14 *†ˠ
Carotid MAP (mmHg)	N/A	80 ± 13	77 ± 16	74 ± 17	69 ± 7	68 ± 17	62 ± 11	57 ± 11
Carotid Flow (mL/min)	N/A	442 ± 13	449 ± 29	419 ± 39	397 ± 133	429 ± 123	361 ± 86	294 ± 98
Coronary APV (cm/s)	N/A	28 ± 10	27 ± 12	24 ± 13	16 ± 2	18 ± 5	20 ± 6	19 ± 7
LV ESP (mmHg)	97 ± 17	95 ± 12	91 ± 14	88 ± 13	76 ± 8	73 ± 8	71 ± 6	64 ± 13
LV EDP (mmHg)	5 ± 5	12 ± 7	14 ± 9	12 ± 8	6 ± 8	6 ± 5	5 ± 5	3 ± 5
LV SW (mmHg·mL)	4153 ± 314	2904 ± 815‡	2888 ± 907	2875 ± 908	2295 ± 510‡	2290 ± 346	2012 ± 387	1752 ± 595
LV ESV (mL)	186 ± 16	193 ± 55	198 ± 56	200 ± 55	129 ± 15ˠ	143 ± 16	145 ± 18	146 ± 21
LV EDV (mL)	236 ± 20	240 ± 61	241 ± 66	245 ± 64	165 ± 22ˠ	180 ± 18 *	179 ± 20 *	178 ± 22
RA (mmHg)	4 ± 2	4 ± 2	4 ± 5	3 ± 4	4 ± 2	5 ± 2	5 ± 2	4 ± 1 †
mean PA (mmHg)	22 ± 8	12 ± 1	15 ± 4	14 ± 4	13 ± 2	15 ± 2	17 ± 3	18 ± 5
PCWP (mmHg)	6 ± 1	9 ± 2	10 ± 7	8 ± 6	9 ± 1	8 ± 2	8 ± 1	7 ± 1
TPG (mmHg)	15.8 ± 7.3	3.4 ± 0.9	5.3 ± 3.5	5.8 ± 2.6	3.7 ± 0.9	7.3 ± 2.6	9.7 ± 3.8 *	11 ± 4.7 *†
SVR (Woods units)	21 ± 2.7	14.8 ± 6.1	14.5 ± 5.5	13.3 ± 5.7	11.4 ± 1.5	7.7 ± 2.2*	9 ± 3.2	8 ± 2.6*
Arterial Ea (mmHg/mL)	2.2 ± 0.6	1.3 ± 0.6	1.5 ± 0.6	1.3 ± 0.5	1 ± 0.2	0.8 ± 0.2*	0.9 ± 0.3	1.9 ± 2.8
TD CO (L/min)	4.1 ± 0.6	5.1 ± 1.2	5.1 ± 1.1	5.4 ± 1.7	5.3 ± 0.8	7.2 ± 1.3 *	6.1 ± 1.1	5.1 ± 2.1
PA Ea (mmHg/mL)	0.7 ± 0.3	0.24 ± 0.08	0.34 ± 0.16	0.31 ± 0.12	0.56 ± 0.17	0.57 ± 0.1	0.71 ± 0.17	0.8 ± 0.19
PA Compliance (mL/mmHg)	2.6 ± 1	11.2 ± 7.3	7.2 ± 2.2	7.6 ± 2.4	4.49 ± 3.01	3.72 ± 0.47	3.36 ± 1.01	2.89 ± 0.68 †
PVR (Woods units)	3.8 ± 1.5	0.7 ± 0.2	1.1 ± 0.8	1.2 ± 0.8	0.70 ± 0.12	1.04 ± 0.41	1.73 ± 1.07	2.82 ± 1.93
RA: PCWP	0.6 ± 0.3	0.4 ± 0.18	0.24 ± 0.28	0.27 ± 0.27	0.48 ± 0.18	0.7 ± 0.19	0.7 ± 0.33	0.5 ± 0.11 †

Ao – Aortic, MAP - mean arterial pressure, APV - average peak velocity, ESP – end systolic pressure, EDP – end diastolic pressure, ESV – end systolic volume, SW – stroke work, RA – right atrium, PA – pulmonary artery, PCWP – pulmonary capillary wedge pressure, TPG - transpulmonary pressure gradient, SVR - systemic vascular resistance, Ea - elastance, TD CO – thermodilution cardiac output, PVR – pulmonary vascular resistance

^*p < 0.05 any pump speed vs. baseline

^†p < 0.05 low pump speed vs. high pump speed

^ˠp < 0.05 Thoracic vs. Abdominal at condition

^‡p < 0.05 vs. Control

Aortic pressure, Carotid Pressure, Coronary Flow

Abdominal activation increased distal aortic diastolic pressure at medium and high pump speeds (Baseline: 49 ± 8 mmHg; vs. medium: 67 ± 8 mmHg, p = 0.03; vs. high: 74 ± 8 mmHg, p = 0.04), while generating a trans-aortic gradient (proximal minus distal mean aortic pressures) at all pump speeds (Baseline: −1 ± 2 mmHg; Low: 12 ± 6 mmHg; Med: 78 ± 5 mmHg; High: 80 ± 10 mmHg, p < 0.0003) (Figure 1D). Abdominal activation generated larger trans-aortic gradients than thoracic positioning. Thoracic activation did not change distal aortic pressure but generated a trans-aortic gradient at medium and high pump speeds. No change in proximal aortic or carotid pressures (systolic, diastolic, mean), carotid flow, or coronary flow velocities were observed regardless of pump position or speed.

LV hemodynamics

Abdominal activation increased thermodilution cardiac output (5.3 ± 0.8 vs. 7.2 ± 1.3 L/min, p = 0.046), reduced estimated arterial elastance (LVESP divided by LVSV, ^{9, 10}) (1 ± 0.2 vs. 0.8 ± 0.2 mmHg/mL, p = 0.002), and systemic vascular resistance at low pump speeds (11.4 ± 1.5 vs. 7.7 ± 2.2 woods units, p = 0.006). Thoracic activation had no effect on these three parameters. Neither abdominal nor thoracic positioning changed LV SW, end-systolic, or end-diastolic pressure at any speed. Compared to baseline values, abdominal activation increased LV end-diastolic volume at low and medium speeds. No change in LV volumes was observed with thoracic activation.

RV hemodynamics/Pulmonary Characteristics

Compared to baseline values, abdominal activation increased trans-pulmonary pressure gradients at medium and high speeds (Baseline: 3.7 ± 0.9 mmHg; vs. medium: 9.7 ± 3.8 mmHg, p = 0.044; vs. high 11 ± 4.7 mmHg, p = 0.036), which was driven by trends towards higher mean pulmonary artery pressures and lower pulmonary capillary wedge pressures. Compared to baseline, pulmonary vascular resistance trended towards increased values at medium and high speed, while pulmonary artery compliance was reduced at high speeds. No change in any index of right ventricular or pulmonary artery hemodynamics was observed with thoracic activation.

Discussion

This is the first report to determine that in contrast to thoracic positioning, abdominal positioning of the newest generation Aortix device reduces LV afterload and increases cardiac output at low speeds. We further observed that the Aortix device does not reduce LV stroke work across all speeds or positions and further that at high speeds the Aortix device may increase LV end-diastolic volume and trans-pulmonary gradients, which may reflect increased venous return leading to increased LV volume and pulmonary vascular auto-regulation ¹¹. Importantly, no significant change in coronary or carotid blood flow or pressure was observed across the various speeds in the thoracic or abdominal position. These findings have potentially important implications for the design of early clinical studies by suggesting that device position and speed are major determinants of hemodynamic efficacy. Specifically, our findings suggest that monitoring patients for new or worsening pulmonary hypertension or right ventricular failure during device activation may be important.

The entrainment mechanism underlying the Aortix device is distinct from other acute mechanical circulatory support (AMCS) pumps. Rotational kinetic energy generated by the device creates oscillatory rotational flow of blood in the aorta above the pump and generates a negative pressure head ¹², thereby reducing LV afterload and increasing cardiac output. Based on this principle, we now report that the Aortix device is currently the only pump platform that can create this trans-aortic gradient (Figure 1D). We did not see a significant or consistent reduction in LV SW indicating a limited unloading effect of the pump on the left ventricle.

Another novel, yet paradoxical, finding was that the Aortix device is more effective at low, not high, speed despite creating a smaller trans-aortic gradient. There are several possible explanations for this observation. Since the Aortix impeller does not occupy the full diameter of the descending aorta, pulsatile blood flow moves around the non-pulsatile flow generated by the pump. At higher speeds, as distal pressure increases, turbulence may develop around the pump as laminar flow interacts with pulsatile flow, which may reduce the negative pressure head above the pump. Increased distal pressure may also transmit retrograde across the pump and increase LV afterload enough to attenuate any increase in cardiac output. Another possible explanation is that increases in distal arterial pressures may promote venous return and thereby increase cardiac preload, ultimately diminishing LV output in the setting of LV injury. For these reasons, low speeds may be necessary and sufficient to reduce LV afterload and enhance cardiac output. Future studies are required to confirm these measurements in additional preclinical and clinical models.

There are several limitations to our study. Given the cost of large animal studies, the n-numbers are low and we were unable to study the hemodynamic effects of more prolonged ramping with the device. Thoracic and abdominal groups had differences at baseline; however, all comparisons inter-pump analyses were made comparing relative changes from baseline. Future studies will focus on renal blood flow parameters, urine output and whether alterations in these indices come at the expense of myocardial and/or cerebral perfusion. This study does not definitely determine whether steal may or may not occur by the Aortix device. Additionally, future studies are required to test the Aortix device in models of acute and chronic heart failure. Finally, no direct comparison with existing pump platforms was performed.

Based on our findings, the Aortix device may be better suited as a mechanical approach to reduce LV afterload, increase renal arterial pressure, and increase cardiac output for patients with decompensated heart failure as opposed to patients receiving AMCS for high risk cardiac interventions or cardiogenic shock. In this preclinical model of LV injury, the Aortix device does not appear to unload the LV. Future studies are required to evaluate whether the Aortix device increases renal perfusion and urine output among patients with decompensated heart failure refractory to medical therapy.

Clinical Perspective.

What is new?

• This is the first report comparing the hemodynamic effects of thoracic versus abdominal aortic positioning of a novel percutaneously-delivered, catheter-mounted, axial-flow pump (Aortix, Procyrion Inc) in a preclinical model of LV dysfunction. The Aortix device is an impeller pump that accelerates native aortic flow creating a trans-aortic pressure gradient.

• We introduce that abdominal positioning at low speeds reduces LV afterload and increases cardiac output without reducing coronary or carotid pressure or flow.

• However, no significant or consistent reduction in LV stroke work was observed with pump activation in either thoracic or abdominal positions.

What are the clinical implications?

• At present, no percutaneously delivered continuous flow pump is designed to support patients with heart failure without cardiogenic shock by reducing LV afterload. Abdominal intra-aortic positioning of the Aortix device reduces LV afterload and increases cardiac output.

• Our observations provide key insight into the design of trials studying the clinical utility of the Aortix device by suggesting that abdominal, not thoracic, positioning increases cardiac output.