Evaluation of a new magnetically suspended centrifugal neonatal pump in healthy animals using a veno-venous extracorporeal membrane oxygenation configuration

Abstract

Background

The objective of this animal study was to evaluate the hemodynamic performance of a new centrifugal pump for extra-corporeal membrane oxygenation (ECMO) support in neonates.

Methods

Six healthy swines were supported with veno-venous ECMO with the New Born ECMOLife centrifugal pump (Eurosets, Medolla, Italy) at different flow rates: 0.25, 0.5, 0.6, and 0.8 L/min; three animals were evaluated at low-flows (0.25 and 0.5 L/min) and three at high-flows (0.6 and 0.8 L/min). Each flow was maintained for 4 hours. Blood samples were collected at different time-points. Hematological and biochemical parameters and ECMO parameters [flow, revolutions per minute (RPM), drainage pressure, and the oxygenator pressure drop] were evaluated.

Results

The increase of the pump flow from 0.25 to 0.5 L/min or from 0.6 to 0.8 L/min required significantly higher RPM and produced significantly higher pump pressures [from 0.25 to 0.5 L/min: 1470 (1253–1569) versus 2652 (2589–2750) RPM and 40 (26–57) versus 125 (113–139) mmHg, respectively; p < .0001 for both - from 0.60 to 0.8 L/min: 1950 (1901–2271) versus 2428 (2400–2518) RPM and 66 (62–86) versus 106 (101–113) mmHg, respectively; p < .0001 for both]. Median drainage pressure significantly decreased from −18 (−22; −16) mmHg to −55 (−63; −48) mmHg when the pump flow was increased from 0.25 to 0.5 L/min (p < .0001). When pump flow increased from 0.6 to 0.8 L/min, drainage pressure decreased from −32 (−39; −24) mmHg to −50 (−52; −43) mmHg, (p < .0001). Compared to pre-ECMO values, the median levels of lactate dehydrogenase, d-dimer, hematocrit, and platelet count decreased after ECMO start at all flow rates, probably due to hemodilution. Plasma-free hemoglobin, instead, showed a modest increase compared to pre-ECMO values during all experiments at different pump flow rates. However, these changes were not clinically relevant.

Conclusions

In this animal study, the “New Born ECMOLife” centrifugal pump showed good hemodynamic performance. Long-term studies are needed to evaluate biocompatibility of this new ECMO pump.

Introduction

Pediatric extracorporeal membrane oxygenation (ECMO) is a supportive strategy used to manage patients with refractory cardiac, respiratory, or cardiorespiratory failure despite maximal medical therapy. ¹ ECMO may serve as a bridge to organ recovery, to decision in life-threatening conditions, to long-term mechanical support, or as a bridge to heart, lung–heart or lung transplantation. To improve the hemodynamic performance while reducing the risk of hemolysis and thrombus formation during ECMO, newer pediatric pumps [e.g., the PediVas (Abbott, Chicago, IL, USA),^2,3 the RotaFlow (Getinge, Rastatt, Germany), ⁴ the Quantum CP22 (Spectrum Medical, Mirandola, Italy) ⁵ and the the Xenios DP3 (Xenios AG, Heilbronn, Germany) ⁶ ] have been developed or adult pumps have been adapted for pediatric use and released on the market over the past two decades.

The PediVas (Abbott, Chicago, IL, USA) is a neonatal/pediatric magnetically suspended centrifugal pump with a static prime of 14 mL and with pump connectors of ¼ inch.^2,3 The RotaFlow (Getinge, Rastatt, Germany) is an adult centrifugal pump used also in neonates/children and is constituted of a magnetically stabilized rotor on a monopivot. This pump has a static prime of 32 mL and requires 3/8 to ¼ inch reducing connectors when used in neonatal circuits. ⁴ The Quantum CP22 centrifugal pump (Spectrum Medical, Mirandola, Italy) is an adult centrifugal pump used also in neonates/children and is constituted of a single sapphire bearing structure. This pump has a static prime of 22 mL, and requires 3/8 to ¼ inch reducing connectors when used in neonatal circuits. Data on hemodynamics and biocompatibility are lacking for neonates and children. ⁵ The Xenios DP3 pump (Xenios AG, Heilbronn, Germany) is a hybrid pump that unifies the benefits of both radial and axial pumps. ⁶ Static prime is 16 mL and the impeller resides on a ceramic single-ball bearing. When used in neonatal circuits, it is customized with ¼ inch connectors.

Based on these characteristics, caution should be used when using these pumps at low blood flow rates because of the risk of recirculation, blood stagnation in the pump shroud, and hemolysis.^2–9 Considering all these limitations, the aim of this animal experiment was to evaluate the hemodynamic performance of a newly developed magnetically suspended centrifugal pump for neonates and infants.

Methods

The new born ECMO Life centrifugal pump

The New Born ECMOLife (New Born ECMOLife, Eurosets, Medolla, Italy) is a magnetically levitated centrifugal pump comprised of a rotating impeller with six blades encased in a polycarbonate housing. All the compnents are coated with phosphorylcholine (Figure 1). The pump prime is 22 mL and it can provide blood flows ranging from 0 to 3 L/min, with a maximum of 4500 RPM. It is designed with ¼ inch inlet/outlet connectors and uses the same drive unit and console of the adult ECMOLife device. ¹⁰ This neonatal pump is paired with a neonatal oxygenator (ECMO Newborn, Eurosets, Medolla, Italy) and is integrated into a dedicated phosphorylcholine-coated PVC ECMO circuit (¼ inch). The circuit is totally integrated with pressure sensors that monitor drainage pressure, as well as pre- and post-oxygenator pressures. The total circuit length is 100 cm to minimize hemodilution. The ECMOLife console also includes probes to measure the delivered ECMO flow, pre-oxygenator oxygen saturation, the hemoglobin concentration and to detect gaseous microemboli.

Figure 1.

Schematic representation of the experimental setting (a–b). New Born ECMOLife centrifugal pump (c). FiO₂: Fraction of inspired oxygen; PEEP: positive end expiratory pressure, RR: respiratory rate, Vt: tidal volume, SpO₂: oxygen saturation.

Animal preparation

Due to supply difficulties of newborn swines at the time of the experiment, healthy young (4 months old) swines were used for this experiment. Considering the high reproducibility of the study and the request to avoid unnecessary use of animals from the Institutional Review Board for Animal Care (IRBAC), six healthy pigs (Sus Scrofa Domesticus) were included in this study. The study received the approval by the Institutional Review Board of the Free University of Brussels (Belgium) (protocol number: 826 N).

On the day of the experiment, the animal underwent a 12-h fast with free access to water. Anesthesia was induced with a combined intramuscular injection of midazolam (1 mg/kg, Mylan, Auckland, New Zealand) and ketamine (20 mg/kg, Dechra, Lille, Belgium). The animal was placed in the supine position and prepared for intubation. A peripheral venous catheter (18-gauge) in a vein of the ear was inserted. A femoral arterial catheter [4.5 French (Fr), Vygon, Ecouen, France] was inserted in a femoral artery and connected to a pressure transducer (True Wave, Edwards, CA, USA) for invasive arterial pressure monitoring and blood gas analysis. Throughout the experiment, the animal was monitored with a continuous electrocardiogram. Data were recorded at different time points and stored in an electronic report form.

After a sequential intravenous injection of 1 mg of atropine sulfate (Sterop, Anderlecht, Belgium), 20 µg/kg of morphine hydrochloride (Sterop, Anderlecht, Belgium), 1 mg/kg of propofol 2% (Fresenius Propoven 2%, Belgium) and 1 mg/kg of rocuronium (Esmeron, MSD, Kenilworth, NJ, USA), an 8 mm endotracheal tube (Medtronic, Minneapolis, MN, USA) was placed, and mechanical ventilation was started in volume controlled mode (Primus, Drägerwerk AG and Co, Frankfurt, Germany) with a tidal volume of 8 mL/kg, 5 cm H₂O of positive end-expiratory pressure (PEEP), a fraction of inspired oxygen (FiO₂) of 1.0, and an inspiratory to expiratory time ratio of 1:2. Sevoflurane (Sevoflo, Abbott, Abbott Park, IL, USA) was used for maintenance of anesthesia at an expiratory percentage between 1.8 and 2.5%.

Ventilation parameters were adjusted to ensure an end-tidal CO₂ between 35 and 45 mmHg and a PaO₂ > 70 mmHg, using the minimum required FiO₂. Positive end-expiratory pressure was arbitrarily set at 5 cm H₂O. A continuous infusion of rocuronium (2–4 mg/kg/h) and morphine hydrochloride (0.2–0.4 mg/kg/h) was maintained, and a balanced crystalloid sultion (Plasmalyte, Baxter, Belgium) was intravenously administered at a rate of 20–25 mL/Kg/h. A continuous infusion of 20% glucose solution (10 mL/h) was started to avoid hypoglycemia. A supra-pubic 14 Fr Foley catheter was surgically inserted to measure the urine output. Under ultrasound guidance, a 5 Fr triple lumen central venous catheter (Arrow International, Reading, PA, USA) was placed in the left external jugular vein and all drug infusions were transferred to the distal line. Central venous pressure was maintained constant between 8 and 10 cm H₂O throughout the experiment. Before the beginning of the experiment, 2 g of amoxicillin-clavulanate were administered to the animal.

To provide blood flows of 0.25 and 0.5 L/min, a 8 Fr venous cannula (REVAS, Free Life Medical GmbH, Aachen, Germany) was inserted in the right jugular vein and a 8 Fr arterial cannula (REVAS, Free Life Medical GmbH, Aachen, Germany) was inserted in the right femoral vein. To provide blood flows of 0.6 and 0.8 L/min, a 10 Fr drainage cannula (REVAS, Free Life Medical GmbH, Aachen, Germany) was inserted into the right jugular vein and a 10 Fr return cannula (REVAS, Free Life Medical GmbH, Aachen, Germany) was inserted into the right femoral vein. All cannuale were RHEOPAX^TM coated (heparin-based coating) and approved for use up to 30 days, and cannualtions were performed under ultrasound guidance.

Prior to cannulation, systemic unfractioned heparin (Liquemin; F. Hoffmann-La Roche & Co., Basel, Switzerland) was administered intravenously at a dose of 100 UI/kg body weight. The priming volume of 350 mL was prepared using crystalloids (Ringer’s Lactate, Baxter, Lessines, Belgium) and the following medications: (a) 150 UI of heparin and (b) 10 mEq of sodium bicarbonate. After cannulation, veno-venous ECMO support was established and mechanical ventilation settings (FiO₂, respiratory rate, and tidal volume) were reduced to maintain normal gas exchange levels for the animal. The blood thermo-regulation was maintained at 37°C using the ECMOLife Heater Unit (Eurosets, Medolla, Italy).

Following the preparation, the animal was stabilized on the surgical table. For anticoagulation, unfractioned heparin infusion was adjusted to achieve an activated clotting time (ACT) between 180 and 220 s, monitored with a dedicated point-of-care device (i-STAT, Kaolin ACT, Abbott Park, IL, USA). Hypotension during the experiments was managed with boluses of 20 mL/kg of balanced crystalloids (Ringer’s Lactate, Baxter, Lessines, Belgium).

Study protocol

Four different pump flow rates were assessed during the perfusion protocol: 0.25, 0.5, 0.6, and 0.8 L/min. The testing involved three swines for the first two flows (low-flow group: 0.25 and 0.5 L/min) and three swines for the other two flows (high-flow group: 0.6 and 0.8 L/min). Each flow rate was maintained for 4 h (4 h for the first flow and 4 h for the second flow), thus, the total duration of the experiment was approximately 8 h. Blood samples of 10 mL each were collected just before the initiation of ECMO (baseline), 20 min after ECMO initiation (T0), at the first and fourth hour of perfusion (T1 and T2) using the first pump flow, and 20 min after the start of the second pump flow (T3), 1 h after the start of the second pump flow (T4), and 4 h after the start of the second pump flow (T5). The collected samples were analyzed to evaluate hematological and biochemical parameters: hematocrit, platelet count, lactate dehydrogenase (LDH), d-dimers and plasma-free hemoglobin. The hemodynamic pump performances were evaluated by monitoring the median RPM, pump pressure, drainage pressure, and the pressure drop across the oxygenator at different flow rates.

Statistical analyses

Quantitative parameters are expressed as median and ranges. The Mann-Whitney test was used to compare hemodynamic parameters at different flow (0.5 vs 0.25 L/min; 0.8 vs 0.6 L/min; 0.6 vs 0.5 L/min). The Friedman test was used to compare the hemodynamic and biologic parameters at different time points and at different flow rates. A p value < 0.05 were considered statistically significant. Prism (Graphpad, San Diego, USA) was used to performe the analyses.

Results

The total duration of experiments were 8.4 (8.2–8.5) hours. Baseline weight of the animals was 34 (20–36) kg, static compliance of the respiratory system 33 (26–35) cm H₂O/mL and driving pressure 12 (10–14) cm H₂O. After the ECMO start, several parameters, including respiraratory rate, tidal volume, minute ventilation and respiratory system compliance were significantly reduced.

Hemodynamic pump performance

At each pump flow tested under identical experimental settings, RPM and pump pressures did not change signifincantly over time. Increase of the pump flow from 0.25 to 0.5 L/min or from 0.6 to 0.8 L/min required significantly higher RPM and produced significantly higher pump pressures [from 0.25 to 0.5 L/min: 1470 (1253–1569) versus 2652 (2589–2750) RPM and 40 (26–57) versus 125 (113–139) mmHg, respectively; p < .0001 for both - from 0.60 to 0.8 L/min: 1950 (1901–2271) versus 2428 (2400–2518) RPM and 66 (62–86) versus 106 (101–113) mmHg, respectively; p < .0001 for both]. Notably, median RPM and pump pressures were significantly higher at 0.5 L/min using a 8 Fr return cannula rather than at 0.6 (p < .0001) and at 0.8 L/min (p < .0001) using a 10 Fr return cannula (Figures 2 and 3).

Figure 2.

Hemodynamic performances (Revolution per minute) of the New Born ECMOLife centrifugal pump during the 8 hours experiment at different flow rates.

Figure 3.

Hemodynamic performances (Pump Pressure) of the New Born ECMOLife centrifugal pump during the 8 hours experiment at different flow rates.

Under identical experimental settings at each pump flow tested, the median drainage pressure and the oxygenator pressure drop did not change signifincantly over time. Median drainage pressure significantly decreased from −18 (−22; −16) mmHg to −55 (−63; −48) mmHg when the pump flow was increased from 0.25 to 0.5 L/min (p < .0001). When pump flow increased from 0.6 to 0.8 L/min, the drainage pressure decreased from −32 (−39; −24) mmHg to −50 (−52; −43) mmHg, (p < .0001). Median drainage pressure at 0.6 L/min was significantly higher (10 Fr drainage cannula) than at 0.5 L/min (8 Fr drainage cannula) (p < .0001; Figure 4). Median oxygenator pressure drop significantly increased from 5 (3–6) mmHg to 14 (7–19) mmHg when the pump flow was increased from 0.25 to 0.5 L/min (p < .0001). Same pattern was seen when the flow was increased from 0.6 to 0.8 L/min [11 (10–13) mmHg versus 15 (10–19) mmHg; (p < .01)] – (Figures 4–5).

Figure 4.

Hemodynamic performances (Drainage Pressure) of the New Born ECMOLife centrifugal pump during the 8 hours experiment at different flow rates.

Figure 5.

Hemodynamic performances (Pressure Drop oxygenator) of the New Born ECMOLife centrifugal pump during the 8 hours experiment at different flow rates.

The median levels of LDH, d-dimers, hematocrit and platelets variably changed during the experiments both in the low-flow group (Figure 6(a), (c), (e) and (g)) and the high-flow group (Figure 6(b), (d), (f) and (h)); nevertheless, all these variables decreased, but not significantly at the end of each experiment compared with the pre-ECMO values (baseline vs T5); median plasma free hemoglobin levels, instead, modestly increased when compared to the pre-ECMO values (baseline vs T5) (low-flow group: p: 0.03; high-flow group: p: 0.04; Figure 6(i) and (l)).

Figure 6.

Markers of hemocompatibility during the 8 hours experiment at different flow rates.

Discussion

This observational animal study reported that the New Born ECMOLife centrifugal pump has good hemodynamic performance. In particular, the study showed that: (a) RPM and pump pressures remained stable over time at each chosen flow rate, and that (b) the pump pressures significantly increased with increasing flow rates.

Some of these hemodynamic findings can be explained by considering the diameters of the drainage and return cannulae. In this observational animal study, an 8 Fr cannula was used to test pump flows between 0.25 and 0.5 L/min, while a 10 Fr cannula was used to test pump flows between 0.6 and 0.8 L/min. Not surprisingly, it was observed that the cannula size is the major contributor to the flow resistance. Median RPM and pump pressures were significantly higher at 0.5 L/min (using an 8 Fr return cannula) compared with 0.6 L/min (using a 10 Fr return cannula), suggesting that part of the pump energy was dissipated to overcome the resistance of the 8 Fr return cannula. Similarly, the drainage pressure, which normally becomes more negative when increasing the pump flow, was less negative at 0.6 L/min with a 10 Fr drainage cannula rather than at 0.5 L/min with a 8 Fr drainage cannula.

These preliminary findings could be potentially used to evaluate the hemodynamic performances of the New Born ECMOLife in comparison with other ECMO pumps (e.g., the PediVas, Rotaflow, and Xenios DP3) used in children. In our animal study, to maintain a flow rate of 0.6 L/min with a 10 Fr drainage/return cannula, a median value of 1950 RPM (1901–2271) was required, while to maintain a flow rate of 0.8 L/min with a 10 Fr drainage/return cannula, a median value of 2428 RPM (2400–2518) was required. Wang et al., ² using a mock circuit with a 10 Fr return cannula, observed that the PediVas required 2800 RPM to maintain a pump flow of 0.6 L/min and 3200 RPM to maintain a pump flow of 0.8 L/min, while the Rotaflow required lower RPM, similar to the ones reported in our experiment with the New Born ECMOLife centrifugal pump. To provide the same pump flow rates, instead, the Xenios DP3 required higher RPM (∼4500 RPM). ⁶ For the best of our knowledge, no hemodynamic data have been presented in the literature for any comparison with the Quantum CP22.

When analyzing these data, even though performed in different experimental settings (mock-circuit vs animals), we can speculate that both the PediVas and the New Born ECMO pump have similar hemodynamic performances. Nevertheless, the PediVas may require higher pump speed than the New Born ECMOLife to achieve the same blood flow in part because of its lower pump prime (14 mL vs 22 mL). On the other hand, the New Born ECMOLife seems to have better hemodynamic performances than the Rotaflow, despite its lower pump prime (22 mL vs 32 mL). Comparing the hemodynamic performances of the New Born ECMOLife with those of the Xenios DP3 is more challenging because the Xenios DP3 is a hybrid pump and requires higher RPMs to achieve the same pump flows of 0.6–0.8 L/min, which may potentially increase the risk of hemolysis in long-term experiments.

In our study the median levels of LDH, d-dimers, hematocrit and platelets variably changed during the experiments both in the low-flow group and the high-flow group; nevertheless, all these variables decreased, but not significantly at the end of each experiment compared with the pre-ECMO values. The plasma free hemoglobin levels modestly increased (<1 mg/dL) at the end of the experiments compared to the pre-ECMO values, however due to the hemodilution effect caused by the ECMO priming and to the different quantity of fluids used in each animal for maintenance (wide range of weight: 20–36 kg), we cannot draw any conclusion on hemocompatibility.

This study has some limitations that need to be acknowledged. First, due to ethical constraints imposed by our ethical committee, only six young swines were included in the experiments, thus, other pediatric pump flow rates such as 1–1.5 L/min could not be tested. Second, the use of “young” (4 months) animals instead of the newborn ones may have biased our results since young animals have different rheologic characteristics (e.g., lower hematocrit) compared with the newborn ones. Third, these animals despite similar ages (4 months) had variable weights (20–36 kg) and this may have increased the hemodilution effect due the amount of fluid received by the animals for mainatenance. Fourth, the use of “healthy” animals may have important clinical implications (hemodynamic stability) which makes translation of our results difficult in severely injured neonates. Fifth, these experiments were limited to a duration of 8 hours, which may not be sufficient to evaluate the hemocompatibility of any pump. Furthermore, we did not measure the plasma concentration haptoglobin. This scavenging protein has a variable plasma concentration (lower in neonates than in adults) and has an important role in reducing the level of plasma free hemoblobin to prevent the development of acute kidney injury.^11,12 Sixth, we were unable to perform a cost analysis of the New Born ECMOLife compared to the existing centrifugal pumps used in neonatal ECMO (PediVas, Rotaflow, Quantum CP22 and Xenios DP3) since the New Born ECMOLife is not yet commercially available. Nonetheless, we believe that the New Born ECMOLife and its associated console should be compared with third-generation pumps such as the CardioHelp (Getinge, Rastatt, Germany) or the Xenios DP3 customized with the MiniLung petite kit (Xenios, Heilbronn, Germany). Both these pumps have flow and pressure sensors integrated into the circuit and provide real-time monitoring of key hemodynamic parameters. However, while the Cardiohelp circuit (HLS Set Advanced 5.0) is customized only with 3/8-inch tubings and is paired with a 1.3 m² oxygenator, both the New Born ECMOLife and the Xenios DP3 centrufugal pumps are customized with ¼ inch tubing and incorporate a neonatal oxygenator. These technical characteristics may show a potential benefit in reducing the hemodilution effect. This difference in tubing size and oxygenator surface area may have significant implications regarding hemodilution and anticoagulation requirements. Overall, further studies are needed to evaluate the clinical efficacy, safety, and cost-effectiveness of the New Born ECMOLife compared to existing ECMO pumps.

Conclusions

This animal study demonstrated that the New Born ECMOLife centrifugal pump exhibits favorable hemodynamic performance in this experimental setting. More extensive investigations are needed to fully ascertain the hemocompatibility of this pump in long-term experiments.