Use of a Pulsatile Ventricular Assist Device (Berlin Heart EXCOR) and an Interventional Lung Assist Device (Novalung) in an Animal Model

Abstract

Pulsatile ventricular assist devices (VADs) are used in pediatric patients mainly as a bridge to heart transplantation. If severe respiratory failure occurs, extracorporeal membrane oxygenation (ECMO) is currently the treatment of choice. ECMO has the potential for severe complications. Interventional lung assist (iLA) devices, e.g., the Novalung, are used in patients with isolated lung failure. This study aimed to show the feasibility of the combined use of the EXCOR VAD (10 ml and 30 ml blood pumps) and the Novalung. There were two separate experiments within this study. First, a bench test was carried out to analyze pressure and flow through both devices. Second, 10 kg and 30 kg pigs received support with the VAD and iLA in series. Pressures and flow were measured systemically before and after the iLA. Flow was unaffected by the iLA. The mean arterial pressure was reduced (mean of 13 mm Hg) by the iLA. There were no obvious difficulties observed within the interaction of VAD and iLA. The combined use of both devices is feasible and able to provide sufficient perfusion pressures. Oxygenation and CO₂ clearance was effectively supported by the iLA. Patients with myocardial and respiratory failure may benefit from this setup.

Ventricular assist devices (VAD) have been successfully used for adult and pediatric patients with myocardial failure as a bridge to cardiac transplantation.¹ A number of patients also suffer from respiratory failure. Instead of using extracorporeal membrane oxygenation (ECMO), this study combines a pulsatile VAD (Berlin Heart EXCOR, Berlin Heart GmbH, Germany) and an interventional lung assist device (iLA, Novalung, GmbH, Hechingen, Germany). Lung assist devices may improve outcome and recovery in respiratory failure and are used predominantly in adults and in larger children (>20 kg) with respiratory failure. The Novalung has been used in more than 3,500 patients and multiple cases have been reported in the literature.^2–5 This study was performed to demonstrate the feasibility of the combined use of both devices.

Materials and Methods

Bench Testing

A bench test was carried out to address technical concerns. The EXCOR (Berlin Heart GmbH, Berlin, Germany) and the Novalung (Novalung GmbH, Hechingen, Germany) were assembled in series (Figure 1). First, a 10 ml polyurethane (PU) valve blood pump was connected to a 22 cm long silicon tubing, inner diameter 6 mm (REHAU AG & Co, Rehau, Germany). The blood pump was controlled by a pneumatic external driving unit (IKUS, Berlin Heart GmbH, Berlin, Germany). The driving pressure was set at 250/−85 mm Hg, the systolic time at 40% and the frequency was increased from 40/min to 150/min. The systemic pressure was adjusted to 100 mm Hg with a filling pressure of 12 mm Hg. The flow was measured with an HT 110 bypass flowmeter (Transonic Systems Inc., Ithaca, NY) at the outlet cannula.

Figure 1.: Bench test set-up.

Second, a 30 ml PU valve blood pump was connected to a 28 cm long silicon tubing, inner diameter 9 mm (REHAU AG & Co, Rehau, Germany). The driving pressure was set at 285/−90 mm Hg, the systolic time at 40% and the frequency was increased from 40/min to 150/min with the IKUS driving unit. The systemic pressure was adjusted to 120 mm Hg with a filling pressure of 12 mm Hg. The flow was measured with an HT 110 bypass flowmeter (Transonic Systems Inc., Ithaca, NY) at the outlet cannula.

Animal Model

The study was approved by the local animal research ethics board. Three 30 kg and three 10 kg pigs were premedicated following the standard animal laboratory protocol with 0.2 ml/kg Akmezine intramuscularly (Ketamine, Atravet, Atropine). The induction was started with Isoflurane 4% and 2 liters of oxygen by facemask till the loss of consciousness and shallow breathing was observed and then switched to Isoflurane 2% for maintenance. The pigs were intubated under vision and paralyzed with a continuous intravenous infusion of Pancuronium (0.2 mg/kg/h).

A central venous line was placed in the internal jugular vein and an arterial line was in the carotid artery by neck cut down. After systemic heparinization (400 U/kg), the pigs were cannulated for cardiopulmonary bypass (CPB). A right-angled venous cannula (24 Fr for 30 kg pigs, 20 Fr for 10 kg pigs) was inserted into the right atrium (DLP, Medtronic, Inc., Minneapolis, MN). A straight arterial cannula (Jostra, Maquet GmbH & Co. KG, Rastatt, Germany) was placed in the right femoral artery (14 Fr for 30 kg pigs; 10 or 12 Fr for 10 kg pigs). The arterial cannula was placed in the femoral artery rather than the ascending aorta, because placement of both the CPB cannula and Berlin Heart cannula in the same area of the pig aorta is technically difficult due to the very short ascending aorta.

A 30 ml VAD pump was used for the 30 kg pigs and a 10 ml VAD pump for the 10 kg pigs. Figure 2 displays the experimental setup. Curved tubing was re-enforced to prevent kinking when the VAD/iLA was placed beside the pig’s body. The iLA was kept below heart level to prevent air trapping. The VAD was connected to the pneumatic driving unit (IKUS). The systolic driving pressure was set at 200–250 mm Hg and the diastolic pressure at −30 to −50 mm Hg. The systolic time was set at 40% and the frequency was adjusted between 60/min and 100/min as needed. The bypass flow was stopped and the arterial CBP cannula left to replace volume (from suctioning and for blood transfusion). The venous cannula was disconnected.

Figure 2.: Animal experiment set-up.

Flow measurements were performed with the HT 110 bypass flowmeter (Transonic Systems Inc., Ithaca, NY) before and after the iLA. Pressures before and after the iLA as well as systemic pressure were measured using standard transducers. Values were recorded before going on CPB (systemic) as well as after 1 and 2 hours on VAD/iLA. Blood gas, hemoglobin, platelet count and plasma free hemoglobin were measured before and during CPB and on VAD/iLA. Ventilator settings were changed to simulate respiratory insufficiency, aiming for permissive hypercapnia. The respiratory rate on the ventilator was reduced to 3–5 breath/min until the endexpiratory CO₂ was >65 mm Hg. A blood gas analysis was performed to confirm this data.

Results

Bench Testing

The bench results with the 10 ml pump in combination with the Novalung showed a linear performance up to a frequency of 130/min. There was minimal backflow in systole. At a frequency of 100/min a mean flow of 0.77 L/min was measured for the 10 ml PU pumps and was slightly lower than expected. The 30 ml pump in combination with the Novalung showed a linear performance up to a frequency of 110/min. There was no significant backflow. At a frequency of 90/min a mean flow of 2.7 L/min was measured for the 30 ml PU pumps (Figure 3A–D). There were no obvious difficulties observed within the interaction of the pneumatic driving unit, the blood pumps, and the iLA.

Figure 3.: A: Flow and driving pressure over time, 10 ml polyurethane blood pumps. B: Flow pattern with increasing pumping frequency, 10 ml polyurethane blood pumps. C: Flow and driving pressure over time, 30 ml polyurethane blood pumps. D: Flow pattern with increasing pumping frequency, 30 ml polyurethane blood pumps.

Animal Model

Three 30 kg pigs and three 10 kg pigs were included in our study. An additional three pigs served as blood donors. There were no technical difficulties assembling the VAD cannulae, the blood pump, and the iLA together. The average time from going on bypass to going on VAD was 44 min (30–55). The combination of VAD and iLA was observed over a period of 165 min (100–225). The mean flow was 2 L/min (2.0–2.1) in the 30 kg pigs with the 30 ml blood pump and 1 L/min (0.98–1.1) in the 10 kg pigs with the 10 ml blood pump (Table 1). This indicates that there was no change in flow pre and post iLA.

Table 1: Pressure/Flow Characteristics and Hemoglobin/Free Hemoglobin, Thrombocytes on Cardiopulmonary Bypasse and Ventricular Assist Device/Interventinal Lung Assist

To ensure proper function of the iLA, the mean arterial pressure should be at least 60 mm Hg and should not exceed 200 mm Hg. The pressures during the experiment were within these limits. The mean pressure (MP) preiLA and postiLA was decreased by a mean of 13.5 mm Hg (range, 1–25 mm Hg). Figure 4 compares the MP at three time points: the systemic arterial pressure before initiation of VAD/iLA as well as 1 and 2 hours after initiation of the combination of VAD/iLA. The systolic pressure was more affected by the iLA membrane (Table 1). The pressure difference (systolic pre/post) was considerably more pronounced (mean 30 mm Hg) with a substantial interanimal variation (range, 5–116 mm Hg). There were negative diastolic values recorded before the iLA. As shown in Table 1, the MP difference before and after the membrane was 25 mm Hg (range, 14–48).

Figure 4.: Changes in mean arterial pressure, mean pressure before and after interventional lung assist device.

A reduced respiratory function was simulated by decreasing the ventilation and the fraction of inspired oxygen (FiO₂). A rise of endexpiratory CO₂ was subsequently tolerated to levels > 65 mm Hg. The hypercapnia was confirmed by arterial blood gas analysis (paCO₂). Before the oxygen flow to the iLA was initiated. The iLA was able to effectively improve CO₂ clearance as the paCO₂ dropped by a mean of 20 mm Hg with 3l oxygen flow to the iLA (before 45–62 mm Hg, after initiation 18–28 mm Hg) exchange.

Plasma free hemoglobin was measured as it serves as a marker to estimate hemolysis during CPB or while being on mechanical circulatory support. It was above 300 mg/dl (normal value for pigs: 5–15 mg/dl) in 3 pigs after CPB and at the initiation of the VAD and iLA, followed by a plateau without significant increase or a downward trend over the course of the experiment (Figure 5).

Figure 5.: Changes in free hemoglobin before and during ventricular assist device/interventional lung assist device.

At the end of each experiment the animal was euthanized as per local animal laboratory standards, using Pentobarbital 80 mg/kg.

Discussion

The results of our study demonstrate that the combined use of a pulsatile VAD with the iLA is feasible. Both devices have been successfully used as a single application. The Novalung is a pumpless device with a low resistance diffusion membrane that can effectively remove accumulated CO₂ in respiratory failure. The device has been inserted in many different clinical settings to support pulmonary function.^3,4,6 Early treatment with an iLA and the use of different ventilation strategies (e.g., high frequency oxygenation) in adult (or acute) respiratory distress syndrome (ARDS) could help to stabilize the patient faster and improve gas exchange.⁷ We were able to show that hemodynamic support and lung rest with CO₂ removal and oxygenation was achieved. If clinically required, an additional right sided VAD could be also implanted. As a modification, the artificial lung could also be placed on the right side of the circulation, downstream from the right-sided pump which is a more complex set-up.

The suggested combination of two assist devices would serve patients with combined respiratory and hemodynamic failure. Most of these patients are currently treated using ECMO, which is laborious, requires 24 hour attendance of a specialist, and has the potential for serious complications.^8–11 The continuous application of blood products during ECMO support carries an additional risk, and previous studies have shown that less blood products are required when using a pulsatile VAD.¹² Cerebral sequelae were frequently found in children treated with ECMO, and the neurological outcome varies from mild cognitive deficit to cerebral palsy.¹³ These complications are less frequently seen in patients on VAD or iLA support,¹⁴ especially as the incidence of complications for the iLA is mostly due to ischemia in a lower limb,¹⁵ which would not be the case in the model presented.

The Novalung has been approved for temporary use in several European countries (e.g., Germany, Spain, Austria, and United Kingdom) and Canada. In the United States, the Federal Drug Administration granted approval for 6 hours postoperatively; however, the use of the device is essentially off-label.

Patients on ECMO and a VAD usually need a continuous infusion of unfractionated heparin, but bleeding complications and the requirement for blood products seems to be more frequent for patients on ECMO. The iLA has a heparin coated membrane and, therefore, needs less systemic anticoagulation with an activated partial thromboplastin time of 50 seconds considered to be sufficient. ECMO requires cannulation of the neck vessels (carotid artery and internal jugular vein) whereas the combination of pulsatile VAD and iLA entails central cannulation via sternotomy. The VAD/iLA combination has a shorter length of tubing which may imply less heat loss and also less inflammatory response; however, currently there are no data available to support this hypothesis. The filling volume of the Novalung is 240 ml which may possibly have a negative impact on the use in smaller children with a lower circulating blood volume.

The plateau or downward trend in plasma free hemoglobin over the course of the experiment indicates that there was no additional hemolysis caused by the VAD/LA combination. We measured the pressures before and after the iLA membrane and demonstrated that it is possible to get sufficient perfusion pressures. The mean arterial pressure was only reduced by 13 mm Hg using a standard adult size device. We found the iLA in our setting very effective even with minimal oxygen flow necessary to improve gas exchange. In this setup, we evaluated the combined use of VAD and iLA over a maximum of 8 hours, therefore we cannot comment on the safety and effectiveness of prolonged use. A BiVAD configuration with the iLA should be possible as well, but this was not the main focus of our experiment.

Although this preliminary experience does not allow for statistical analysis, the objectives of this study to prove the principle and the feasibility of the combined use of iLA and VAD were achieved and may allow the application to a patient with myocardial and respiratory failure.

Conclusion

The combination of a pulsatile VAD and the iLA in an animal model is feasible. Sufficient perfusion pressures can be achieved with the VAD in combination with a standard size iLA. Oxygenation and CO₂ clearance were effectively achieved by the iLA. Patients with myocardial and respiratory failure may benefit from this setup.

Acknowledgment

Berlin Heart GmbH, Berlin, Germany, supported this study with the supply of the blood pumps and a driving unit (IKUS) which were returned to Berlin Heart after the experiments. Novalung GmbH, Hechingen, Germany, donated the iLAs for single use. Both companies provided a research grant to cover the costs of the animal experiments.

The authors thank Kersten Brandes, MD, for her support in the planning phase of the study.