Evaluation of a New Extracorporeal CO2 Removal Device in an Experimental Setting

Background: Ultra-protective lung ventilation in acute respiratory distress syndrome or early weaning and/or avoidance of mechanical ventilation in decompensated chronic obstructive pulmonary disease may be facilitated by the use of extracorporeal CO2 removal (ECCO2R). We tested the CO2 removal performance of a new ECCO2R (CO2RESET) device in an experimental animal model. Methods: Three healthy pigs were mechanically ventilated and connected to the CO2RESET device (surface area = 1.8 m2, EUROSETS S.r.l., Medolla, Italy). Respiratory settings were adjusted to induce respiratory acidosis with the adjunct of an external source of pure CO2 (target pre membrane lung venous PCO2 (PpreCO2): 80–120 mmHg). The amount of CO2 removed (VCO2, mL/min) by the membrane lung was assessed directly by the ECCO2R device. Results: Before the initiation of ECCO2R, the median PpreCO2 was 102.50 (95.30–118.20) mmHg. Using fixed incremental steps of the sweep gas flow and maintaining a fixed blood flow of 600 mL/min, VCO2 progressively increased from 0 mL/min (gas flow of 0 mL/min) to 170.00 (160.00–200.00) mL/min at a gas flow of 10 L/min. In particular, a high increase of VCO2 was observed increasing the gas flow from 0 to 2 L/min, then, VCO2 tended to progressively achieve a steady-state for higher gas flows. No animal or pump complications were observed. Conclusions: Medium-flow ECCO2R devices with a blood flow of 600 mL/min and a high surface membrane lung (1.8 m2) provided a high VCO2 using moderate sweep gas flows (i.e., >2 L/min) in an experimental swine models with healthy lungs.


Introduction
Significant advancements have been done to understand the feasibility and safety of extracorporeal CO 2 removal (ECCO 2 R) in patients with hypoxemic and/or hypercapnic respiratory failure [1][2][3][4]. ECCO 2 R has been used either to reduce the main components of the mechanical power (i.e., respiratory rate, driving pressure, flow rate and/or positive end expiratory pressure), which may potentially cause ventilator-induced lung injury (VILI) in patients with severe acute respiratory distress syndrome (ARDS) [5], or, to avoid endotracheal intubation and invasive mechanical ventilation in patients failing non-invasive ventilation for acute exacerbations of chronic obstructive pulmonary disease (COPD) or of end-stage respiratory disease awaiting for a lung transplant [3][4][5][6][7]. For these purposes, several devices have been developed (i.e., pumpless arterio-venous or pump-driven venovenous circuits), using a varying range of blood flow (i.e., from 200 to 1800 mL/min), sweep gas flow and different sizes (m 2 ) of membrane lungs. Further, to increase CO 2 removal with minimal blood flow, hybrid techniques have been developed. These techniques, using an acidified dialysate and a hemofilter, may increase the amount of H+ in the blood and consequently the CO 2 content upstream of the membrane lung. So far, hybrid techniques (regional acidification) may increase the amount of CO 2 removed; however, their complexity has limited their clinical use [8,9].
Recently, a multicenter pilot study conducted in patients with moderate ARDS showed that high-flow CO 2 removal devices (i.e., blood flow around 800-1000 mL/min) had fewer hemorrhagic complications and hemolysis than low-flow (i.e., blood flow around 400 mL/min) devices, with a significantly better reduction of PaCO 2 [2]. Animal data, instead, were controversial. Duscio et al. [10] reported very high CO 2 removal (171 mL/min) using a low-flow device (400 mL/min), while Karagiannidis et al. [11,12] showed that only high blood flow rates (>900 mL/min) and adequate membrane lungs (surface area > 1 m 2 ) can effectively correct severe respiratory acidosis. However, both studies converge on the point that the sweep gas flow can increase CO 2 removal only when high blood flow rates are used.
With the present study conducted in healthy pigs, we aimed to describe the CO 2 removal performance and operational characteristics of a new medium-flow ECCO 2 R device, which has been created specifically for CO 2 removal using a fixed amount of blood flow rate, a membrane lung of 1.8 m 2 and different sweep gas flow rates.

Extracorporeal CO 2 Removal Technique
Medium-flow veno-venous ECCO 2 R was performed using the CO 2 RESET device (EUROSETS S.r.l., Medolla, MO, Italy). This device, driven by roller pumps, incorporates both a hemoperfusion membrane and a phosphorylcoline-coated polymethylpentene hollow fiber membrane lung (surface area = 1.8 m 2 ), without an integrated heat exchanger. The membrane lung may be connected either to a sweep gas source of pure oxygen or to a mixture of air/oxygen to provide CO 2 removal. The CO 2 RESET circuit is customized for single-use and can be connected to a wide range of cannulas. In our animal model, the hemoperfusion membrane was not incorporated and the device was used specifically for CO 2 removal ( Figure 1). The CO 2 RESET circuit is customized to receive only 1 4 connectors and connects to either with dual lumen cannulas (13,16,19 French) or two single cannulas.

Animal Preparation
The Institutional Review Board of the Free University of Brussels (Belgium) approved the experimental protocol (number of Ethical Committee approval: 731N). On the day of the experiment, the animal (swine, Sus Scrofa Domesticus) was fasted for 12 h with free access to water. Anesthesia was initiated with a combined intramuscular injection of midazolam (1 mg/kg, Mylan, Auckland New Zeland and ketamine (100 mg/kg, Dechra, Lille, Belgium) administered in the neck and placed in supine position. The animal was monitored with a continuous electrocardiogram, and a peripheral vein (18-gauge) was inserted to provide a continuous infusion of sufentanil citrate (3 µg/kg, Janssen, Beerse, Belgium). A femoral 4.5 French (Fr) arterial catheter (Vygon, Ecouen, France) was inserted in the femoral artery and connected to a pressure transducer (True Wave, Edwards, CA, USA) for invasive arterial pressure monitoring and blood gas analysis (BGA). After a sequential intravenous injection of 1 mg atropine sulfate (Sterop, Anderlecht, Belgium), 3 µg/kg of sufentanil citrate and 1.2 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 controlled volume mode (Primus, Drägerwerk AG and Co, Frankfurt, Germany) with a tidal volume of 8 mL/kg, 5 cmH 2 O of positive end-expiratory pressure (PEEP), fraction of inspired oxygen (FiO 2 ) of 1.0 and an inspiratory to expiratory time ratio of 1 to 2. A 1% mixture of inspired sevofluorane (Sevoflo, Abbott, Abbott Park, IL, USA) was started to achieve an expiratory percentage between 1.2 to 1.6%. Mechanical power of the respiratory system was calculated according to the validated formulas [13].
Membranes 2021, 10, x FOR PEER REVIEW 3 of 11 customized to receive only ¼ connectors and connects to either with dual lumen cannulas (13,16,19 French) or two single cannulas.

Animal Preparation
The Institutional Review Board of the Free University of Brussels (Belgium) approved the experimental protocol (number of Ethical Committee approval: 731N). On the day of the experiment, the animal (swine, Sus Scrofa Domesticus) was fasted for 12 h with free access to water. Anesthesia was initiated with a combined intramuscular injection of midazolam (1 mg/kg, Mylan, Auckland New Zeland and ketamine (100 mg/kg, Dechra, Lille, Belgium) administered in the neck and placed in supine position. The animal was monitored with a continuous electrocardiogram, and a peripheral vein (18-gauge) was inserted to provide a continuous infusion of sufentanil citrate (3 µ g/kg, Janssen, Beerse, Belgium). A femoral 4.5 French (Fr) arterial catheter (Vygon, Ecouen, France) was inserted in the femoral artery and connected to a pressure transducer (True Wave, Edwards, CA, USA) for invasive arterial pressure monitoring and blood gas analysis (BGA). After a sequential intravenous injection of 1 mg atropine sulfate (Sterop, Anderlecht, Belgium), 3 µ g/kg of sufentanil citrate and 1.2 mg/kg of rocuronium Ventilation parameters were subsequently adjusted to ensure an end tidal CO 2 between 35 and 45 mmHg and SpO2 > 96%, using the minimally required FiO 2 . A continuous infusion of rocuronium (2-4 mg/kg/h) and sufentanil citrate (3.5 µg/kg/h) was maintained, and balanced crystalloids (PLASMA-LYTE, Baxter, Lessines, Belgium) were administered at a rate of 300-500 mL/h. A 14 Fr Foley catheter was surgically inserted to measure urine output thorough a midline incision in the lower abdomen, and the parietal layers were sutured separately. Under ultrasound guidance, a 5 Fr triple lumen central venous catheter (Arrow International, Reading, PA, USA) was placed in the right internal jugular vein and drug infusion was transferred to the distal line. For veno-venous ECCO 2 R, a 12 Fr multistage drainage cannula (REVAS, Free Life Medical GmbH, Aachen, Germany) was inserted in the left femoral vein and a 10 Fr return cannula (REVAS, Free Life Medical GmbH, Aachen, Germany) was inserted in the internal left jugular.
At this point, a continuous infusion of propofol (Propovet 2%, Abbott, NJ, USA) at the dose of 2-3 mg/kg was started, allowing for a progressive decrease until complete arrest of the anesthetic gas flow. The respiratory system was then opened with the removal of the soda lime absorber (Drägersorb Free, Drägerwerk AG and Co, Frankfurt, Germany). At the end of the preparation, the animal was proned and stabilized on the surgical table. For anticoagulation, heparin infusion was adjusted to achieve an activate clotting time (ACT) between 180 and 220 s, monitored via a dedicated point of care device (i-STAT, Kaolin ACT, Abbott Park, IL, USA). The estimated CO 2 production in pigs at rest is about 200-250 mL/min [14], which is comparable to an adult human.

Study Design and Experiment Procedure
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), three healthy pigs were included in this study. After oral intubation and induction of respiratory acidosis by reducing mechanical ventilation settings, the pig was connected to the veno-venous ECCO 2 R device (CO 2 RESET, EUROSETS S.r.l., Medolla, MO, Italy).
Respiratory acidosis was induced with a 50% reduction of both the baseline tidal volume (from 8 to 4 mL/kg) and the respiratory frequency, to achieve an end tidal CO 2 (etCO 2 ) between 50 and 60 mmHg ( Figure 2). To ensure a pre-membrane lung CO 2 (P pre CO 2 ) between 80-120 mmHg, an external supplementation of 1 L/min CO 2 was provided to the inspiratory side of the ventilator circuit [15,16]. FiO 2 was adjusted to maintain a SpO 2 > 96%. A pre-membrane lung blood gas analysis (BGA) to control the achievement of this target and an arterial BGA from the animal were undertaken before starting each step of experiment. At this time, the veno-venous ECCO 2 R device was connected with a fixed blood flow rate of 600 mL/min and a sweep gas flow of 0 L/min (i.e., no CO 2 removal capacity). All the experiments started with P pre CO 2 between 80 and 120 mmHg. The experiment included six steps from 0 to 1, 2, 3, 5 and 10 L/min of sweep gas flow, respectively. Blood flow was fixed at 600 mL/min during all the steps. Each step of sweep gas flow was maintained for 30 min, and at the end, a post-membrane lung BGA and an arterial BGA from the animal were sampled and CO 2 elimination (VCO 2 ) was collected. At the end of each step, the sweep gas flows were brought to 0 L/min until the animal reached a P pre CO 2 between 80 and 120 mmHg. Body temperature was maintained stable at 37 • C during the study using a warming blanket (Bair Hugger 3M, Zwijndrecht, Belgium). VCO 2 was calculated directly from the device (multiplying the specific sweep gas flow for the partial pressure of the CO 2 exhaled from the membrane lung) and provided in BTPS (body temperature, pressure, water vapor saturated). Operational characteristics of the ECCO 2 R device, including access, return and pressure drop across the membrane lung, were recorded at each step. Experiments were performed in each pig in a standardized fashion. At the end of the experiment, the animal was sacrificed by injection of 80 mEq of KCl under deep sedation.
CO2 exhaled from the membrane lung) and provided in BTPS (body temperature, pressure, water vapor saturated). Operational characteristics of the ECCO2R device, including access, return and pressure drop across the membrane lung, were recorded at each step. Experiments were performed in each pig in a standardized fashion. At the end of the experiment, the animal was sacrificed by injection of 80 mEq of KCl under deep sedation.

Statistical Analysis
Data are expressed as median and inter-quantile ranges. Descriptive statistical analysis for non-parametric data was performed with Wilcoxon and Friedman test with Dunn's multiple comparisons using GraphPad Prism 8 (San Diego, CA, USA). A p value < 0.05 was defined as statistically significant.

Statistical Analysis
Data are expressed as median and inter-quantile ranges. Descriptive statistical analysis for non-parametric data was performed with Wilcoxon and Friedman test with Dunn's multiple comparisons using GraphPad Prism 8 (San Diego, CA, USA). A p value < 0.05 was defined as statistically significant.

Discussion
The main finding of the present porcine study is that the use of a high-surface membrane lung may substantially impact CO 2 elimination with a stepwise increase of the sweep gas flow rate despite using a fixed relatively blood flow rate (i.e., 600 mL/min). Pressure drops using these settings were within normal ranges. Although the experiment was conducted in a small cohort of healthy animals and with a limited time, no complications or technical issues were reported during the short observational period.
From a technical point of view, CO 2 elimination may be manipulated at the bedside by modifying the blood flow rate, the sweep gas flow or the size of the membrane lung, or by acidifying the blood before the membrane lung (increase of P pre CO 2 ). Animal data conflict with one another. Karagiannidis et al. [11,12] demonstrated, using a porcine model, that a blood flow rate > 1 L/min with a membrane lung with a surface area > 0.8 m 2 may remove the 50% of total CO 2 production and correct severe respiratory acidosis. Contrarily, Duscio et al. [10] reported a very high VCO 2 (171 mL/min) with a low blood flow (400 mL/min) and a high surface membrane lung (1.8 m 2 ).
Our results are in between the previous findings of Karagiannidis et al. [11,12] and Duscio et al. [10]. In our porcine model, using a fixed blood flow rate of 600 mL/min and a high surface membrane lung (1.8 m 2 ), we reported a CO 2 elimination of 170.00 (160.00-200.00) mL/min using 10 L/min of sweep gas starting with a P pre CO 2 around 102.50 (95.30-118.20). Furthermore, CO 2 elimination progressively increased when increasing the gas flow from 0 to 2 L/min and reached a plateau at sweep gas flows higher than 3-4 L/min, in line with other previous studies [11,16].
Using this configuration, our VCO 2 was similar to the ones reported by Karagiannidis et al. [11], which used the same range of P pre CO 2 , higher blood flow rates (1 L/min) and smaller membrane lungs (0.8-1.3 m 2 ), similar to the ones of Duscio et al. [10], which used a lower blood flood (400 mL/min) with a very high membrane surface area (1.8 m 2 ), and higher than the ones reported by Hospach et al. [15], which used a blood flow of 600 mL/min with the PrismaLung (Baxter, Lessines, Belgium) and the A.L.ONE (EU-ROSETS S.r.l., Medolla, MO, Italy) membrane lungs (0.80 and 1.35 m 2 , respectively).
The use of low/medium-flow ECCO 2 R devices with this setting may have some advantages compared to high-flow ECCO 2 R devices. First, low/medium-flow ECCO 2 R are generally driven by roller pumps and have been developed specifically to remove CO 2 [15]. High-flow devices are generally centrifugal pumps designed for higher blood flows and used for extracorporeal membrane oxygenation (ECMO). When these ECMO pumps are adapted to work at a lower blood flow rates and with smaller membrane lungs (neonatal or pediatric), these "adjustments" are not free of risks and may induce platelets activation and/or destruction as well as hemolysis, due to the reduction of the hydraulic efficiency and the increase of both pump recirculation rate and shear stress [17]. Second, circuit priming is fast and similar to the ones used for renal replacement therapy (RRT); furthermore, it does not require dedicated specialists as for ECMO (i.e., perfusionists). Third, low/medium-flow ECCO 2 R devices can integrate other organ support techniques (i.e., RRT) to promptly manage the patients with both acute respiratory failure and acute kidney injury/fluid overload. Together with these potential advantages, some challenges exist for low/medium-flow ECCO 2 R devices. First, to provide adequate level of VCO 2 using a low blood flow, they require a large membrane surface area [10]. The interaction between a large membrane surface area and a low blood flow may induce the development of areas of blood stagnation, increasing the risk of thrombotic complications ("circuit" diffuse intravascular coagulation) and secondary hemolysis. Second, low/medium-flow ECCO 2 R devices require an external heating system to maintain body temperature within normal ranges.
This study presents some limitations. First, due to the ethical concerns raised by our IRBAC, few animals have been included; thus, our data cannot be directly transferred to humans with acute respiratory distress syndrome (ARDS) or decompensated COPD. Further, the blood and the interstitial fluid account only for less than 20% of the total human CO 2 stores that can be mobilized within 48 h [18]. Thus, adding CO 2 with an external source for a limited amount of time is not the most accurate strategy to increase CO 2 storages [18]. However, we adopted this approach, previously used by other authors [15,16] to rapidly increase P pre CO 2 , avoiding the reduction of tidal volume (<4 mL/kg) and respiratory rate to unsafe levels that could have required an adjustment of the ventilator settings.
Second, only one fixed blood flow (600 mL/min) was used in our experiments, although the CO 2 elimination is known to increase with the increase of blood flow; the operating range of CO 2 RESET is wider with a maximum of 800 mL/min. However, our purpose was to maintain a blood flow that was comparable with the ones described in other studies [15,17]. Of note, we preferred not to use lower blood flows (i.e., 400 mL/min) [10], to avoid the use of higher ACT (300 sec) for anticoagulation [1].
Third, our study has been designed to maintain a wide range of high P pre CO 2 and consequently provides high VCO 2 . This wide range, even though not common when using ECCO 2 R to prevent VILI in ARDS patients, has also been chosen to test this new ECCO 2 R device in extreme clinical situations such as near fatal asthma and acute exacerbation of COPD or end-stage lung diseases awaiting a lung transplant [7].
Fourth, even though we did not report any mechanical complications (bleeding, clotting, air embolism, circuit failure or hemolysis) during the three experiments, we cannot draw any clinical conclusions since the duration of our experiments was limited.

Conclusions
Medium-flow ECCO 2 R devices with a blood flow of 600 mL/min and a high surface membrane lung (1.8 m 2 ) provided high VCO 2 using moderate sweep gas flows (i.e., >2 L/min) in an experimental swine model with healthy lungs. Future clinical data from these devices would provide further information on this approach in a human setting.