Alkaline Liquid Ventilation of the Membrane Lung for Extracorporeal Carbon Dioxide Removal (ECCO₂R): In Vitro Study

Abstract

Extracorporeal carbon dioxide removal (ECCO₂R) is a promising strategy to manage acute respiratory failure. We hypothesized that ECCO₂R could be enhanced by ventilating the membrane lung with a sodium hydroxide (NaOH) solution with high CO₂ absorbing capacity. A computed mathematical model was implemented to assess NaOH–CO₂ interactions. Subsequently, we compared NaOH infusion, named “alkaline liquid ventilation”, to conventional oxygen sweeping flows. We built an extracorporeal circuit with two polypropylene membrane lungs, one to remove CO₂ and the other to maintain a constant PCO₂ (60 ± 2 mmHg). The circuit was primed with swine blood. Blood flow was 500 mL × min⁻¹. After testing the safety and feasibility of increasing concentrations of aqueous NaOH (up to 100 mmol × L⁻¹), the CO₂ removal capacity of sweeping oxygen was compared to that of 100 mmol × L⁻¹ NaOH. We performed six experiments to randomly test four sweep flows (100, 250, 500, 1000 mL × min⁻¹) for each fluid plus 10 L × min⁻¹ oxygen. Alkaline liquid ventilation proved to be feasible and safe. No damages or hemolysis were detected. NaOH showed higher CO₂ removal capacity compared to oxygen for flows up to 1 L × min⁻¹. However, the highest CO₂ extraction power exerted by NaOH was comparable to that of 10 L × min⁻¹ oxygen. Further studies with dedicated devices are required to exploit potential clinical applications of alkaline liquid ventilation.

1. Introduction

Extracorporeal carbon dioxide removal (ECCO₂R) clears CO₂ from the blood through an extracorporeal membrane lung (ML). This allows independent modulation of minute ventilation and arterial partial pressure of CO₂ (PaCO₂), which are otherwise physiologically linked [1]. ECCO₂R has been proposed to facilitate ultra-protective ventilation [2,3,4] and to promote non-invasive ventilation [5]. This could be particularly beneficial in patients suffering from respiratory failure, including exacerbations of chronic obstructive pulmonary disease (COPD) [6], acute respiratory distress syndrome (ARDS) [7], and patients awaiting lung transplantation [8]. The amount of CO₂ removed by the extracorporeal support is a crucial determinant of clinical efficacy [9,10]. However, the clinical benefits of ECCO₂R are still under evaluation due to safety concerns, mainly related to hemorrhagic and thrombotic adverse events [9].

Several ECCO₂R devices are clinically available. They are mainly characterized by a low extracorporeal blood flow (i.e., <500 mL × min⁻¹) to achieve minimally invasive approaches [11]. Indeed, although 500 mL of blood contain an amount of CO₂ comparable to the amount of CO₂ produced by the body in one minute ($\dot{V}$CO₂), the relatively low CO₂ transfer efficiency of conventional MLs significantly reduces the efficacy of these strategies [12].

The transmembrane gradient of PCO₂ is the driving force that moves CO₂ from blood to the sweeping gases. However, the use of high sweep gas flows, while maximizing the transmembrane gradient, does not increase CO₂ clearance significantly. Indeed, during ECCO₂R, most of the extracorporeal CO₂ removal capacity is achieved for sweep gas flows below 2 L × min⁻¹ since, at higher flows, the system rapidly loses efficiency [13,14,15,16].

Several ECCO₂R techniques are currently undergoing preclinical evaluations. The main aim is to overcome the present limitations to enhance CO₂ removal [17,18,19,20,21] effectively. To this purpose, our group has achieved high rates of CO₂ removal through acidification of the blood entering the ML [22,23,24,25,26,27]. This strategy reduced dissociated CO₂ (HCO₃⁻) in favor of dissolved CO₂ (PCO₂), thus increasing the efficiency of ECCO₂R. Nevertheless, these approaches are still experimental, mainly due to safety and technical issues [28,29].

In the present study, we hypothesized that extracorporeal CO₂ removal could be enhanced through the ventilation of the ML with a sweep fluid with an extremely high CO₂ absorbing capacity (sodium hydroxide -NaOH- solutions), thereby preserving the transmembrane CO₂ gradient.

Indeed, when a high amount of CO₂ is added to dilute NaOH solutions, carbon dioxide first hydrates to carbonic acid (H₂CO₃), Equation (1), which will subsequently react with NaOH to form sodium bicarbonate (NaHCO₃), Equation (2).

CO₂ + H₂O ↔ H₂CO₃

(1)

NaOH + H₂CO₃ ↔ Na⁺ + HCO₃⁻ + H₂O

(2)

Instead, when CO₂ is added to highly concentrated NaOH solutions, sodium bicarbonate is formed directly, Equation (3), which subsequently forms sodium carbonate, Equation (4).

NaOH + CO₂ ↔ NaHCO₃

(3)

NaHCO₃ + NaOH ↔ Na₂CO₃ + H₂O

(4)

Consequently, highly concentrated NaOH solutions can absorb a conspicuous amount of CO₂ while keeping PCO₂ almost down to zero although the elevated pH of the solution causes safety concerns.

The aim of the present proof-of-principle study was to evaluate in-vitro the feasibility and the CO₂ transfer efficacy of membrane lung ventilation with a NaOH solution. This type of ventilation was named “alkaline liquid ventilation”.

Different concentrations of NaOH were tested and the efficacy and efficiency in CO₂ removal of alkaline liquid ventilation were compared to conventional sweep gas flow.

2. Materials and Methods

An in vitro setting (Figure 1) was built to simulate a patient undergoing extracorporeal CO₂ removal.

A closed-loop circuit was assembled with 3/8 and 1/4 inch polyvinylchloride class IV medical tubes (Medtronic, Minneapolis, MN, USA), one 4 L reservoir (VHK 71000 venous hardshell cardiotomy reservoir, Getinge, Gothenburg, Sweden), two polypropylene oxygenators membrane gas exchangers (Quadrox-i Small adult HMO 50000, Getinge, Gothenburg, Sweden) and one peristaltic pump (Multiflow Roller Pump Module H10 series, Stöckert Shiley, München, Germany).

The circuit was primed with about 3 L of swine blood collected at a local abattoir during usual slaughtering processes in compliance with CE regulations (1069/2009), authorization number 0141051/19 provided by ATS Milano, Regione Lombardia. MultiBic^® solution (Fresenius Medical Care Italia, Palazzo Pignano, Italy) was added to achieve a total volume of about 4 L. Sodium Heparin 25000 I.U. (Pfizer Italia S.r.l, Latina, Italy), anticoagulant-citrate-dextrose ACD 300 mL (Fresenius Kabi Italia, Isola Della Scala, Italy) and cefazolin 1 g (Teva Italia, Milano, Italy) were added to the blood.

The first gas exchanger downstream the reservoir was ventilated with a gas mixture of air and CO₂ to maintain a constant PCO₂ of 60 ± 2 mmHg at the inlet of the second oxygenator throughout all experiments. The second oxygenator was employed to remove CO₂ through either ventilation with oxygen or a continuous infusion of sodium hydroxide (NaOH) solution, “alkaline liquid ventilation”, into the gas side of the membrane lung. Circuit accesses for blood sampling were positioned upstream (PRE) and downstream (POST) of the second oxygenator.

NaOH pellets (Sigma-Aldrich, Merck KGaA, Saint Louis, MO, USA) were diluted in distilled water to achieve the required concentrations. NaOH solutions were stored in disposable parenteral bags (Bertoni Nello S.r.l. Modena, Italy) and infused into the gas inlet port in the gas exchanger using a peristaltic pump (Multiflow Roller Pump Module H10 series, Stöckert Shiley, München, Germany). NaOH exiting the oxygenator was discarded.

The blood temperature was kept stable at 37 °C through heat exchangers connected to the membrane lungs.

The study was divided into four steps: (1) a mathematical modeling of NaOH and CO₂ interactions to evaluate the theoretical basis of the study; (2) a safety and feasibility test to evaluate the effects of increasing NaOH concentrations on the membrane lung integrity and CO₂ removal; (3) an efficiency test to compare the CO₂ removal of similar sweep flows (up to 1 L × min⁻¹) of oxygen vs. NaOH at the concentration selected following the feasibility test; (4) an efficacy test to compare the CO₂ removal of the best liquid ventilation flow, selected from the efficiency test, vs. 10 L × min⁻¹ of oxygen.

All the in-vitro tests were performed with 500 mL × min⁻¹ of blood flow.

2.1. Mathematical Modeling

Theoretical effects of CO₂ absorption by aqueous NaOH were computed solving a system of equations (MATLAB R2018b; The Math Works, Inc, Natick, MA, USA), including standard mass-action, mass-conservation and electroneutrality laws of the involved species: water, NaOH, CO₂ (see the Online Supplement for more details).

We simulated a closed system with aqueous NaOH at varying concentrations (from 0 to 100 by 20 mmol × L⁻¹) in which we introduced CO₂ at different concentrations (from 0 to 100 by 5 mmol × L⁻¹).

Of note, in the present mathematical model of a closed system, total pressure could exceed barometric pressure.

2.2. Definitions and Calculations

Bicarbonate ion concentration ([HCO₃⁻]) was calculated from pH and PCO₂ modifying the Henderson-Hasselbalch equation

$$[{\mathrm{HCO}}_3^{-}]= \mathsf{\alpha } \times { \mathrm{PCO}}_2\times {10}^{\mathrm{pH}-\mathrm{pK}}$$

(5)

where α = 0.0307 mmol × L⁻¹ × mmHg⁻¹ (solubility of CO₂ in plasma) [30,31] and pK = 6.129 (negative logarithm of the equilibrium constant) [31,32,33].

Plasma carbon dioxide content PRE and POST membrane lung (expressed in mmol × L⁻¹) was calculated according to the method published by Douglas et al. [34]:

$$[{\mathrm{TCO}}_2]= \mathsf{\alpha } \times { \mathrm{PCO}}_2\times \left(1+{10}^{\mathrm{pH}-\mathrm{pK}}\right)$$

(6)

Carbon dioxide transfer across the membrane lung, $\dot{V}$CO₂ (in mL × min⁻¹), was calculated from the transmembrane lung TCO₂ difference [35]:

$${\dot{V}\mathrm{CO}}_2=(\left[{\mathrm{TCO}}_{2\mathrm{PRE}}\left]-\right[{\mathrm{TCO}}_{2\mathrm{POST}}]\right)\times \mathrm{blood} \mathrm{flow} \times 25.45$$

(7)

where TCO_2PRE represents CO₂ content before the membrane lung while TCO_2POST is the CO₂ content after the membrane lung, blood flow is measured in L × min⁻¹, and the conversion factor is in mL × mmol⁻¹.

2.3. Safety and Feasibility Test

Possible macroscopic detrimental effects on the membrane lung were evaluated. The effect on CO₂ removal of alkaline liquid ventilation at increasing concentrations of NaOH (10, 30, 60, 90, 100 mmol × L⁻¹) and increasing ventilating flows (100, 250, 500, 1000 mL × min⁻¹) was likewise evaluated. Each combination of NaOH concentration and sweep fluid flow was tested once and for 15 min. At the end of each step, PRE and POST blood samples were collected for blood gas analysis (BGA) (Radiometer abl800 flex, Copenhagen, Denmark).

In addition, the integrity of the oxygenator was evaluated through visual inspection of the membrane lung, evaluation of the presence of blood in the NaOH solution exiting the oxygenator, and through analysis of blood sodium, potassium, and methemoglobin as indirect markers of hemolysis. The time-course of methemoglobin was evaluated at 4 time points (15, 30, 45, and 60 min) while testing aqueous NaOH at different sweep flows during the efficiency and efficacy tests.

The CO₂ removal efficiency was estimated by computing PCO₂ differences across the membrane lung and $\dot{V}$CO₂.

At the end of the feasibility test, we selected the highest NaOH concentration endured by the membrane lung to perform the subsequent efficiency and efficacy tests.

2.4. Efficiency and Efficacy Tests

We performed six experiments with blood from 4 pigs. For each experiment, we tested, in random order, two different sweeping fluids, pure oxygen (FiO₂ equal to 1) and aqueous NaOH at 100 mmol × L⁻¹ (the concentration selected from the feasibility test). Four sweep flows (100, 250, 500, 1000 mL × min⁻¹) for each fluid were randomly tested. We also randomized and tested 10 L/min of oxygen flow. Each combination of sweep fluid and flow was applied once during the single experiment.

The target PRE PCO₂ was 60 ± 2 mmHg.

At the end of each step lasting about 15 min, we collected PRE and POST blood samples for BGA.

CO₂ removal efficiency and efficacy were evaluated from PCO₂ differences across the membrane lung and $\dot{V}$CO₂.

The highest CO₂ removal achieved with alkaline liquid ventilation was compared with the CO₂ removal achieved with conventional gaseous ventilation performed with 10 L × min⁻¹ of oxygen.

2.5. Statistical Analysis

Data are reported as median and interquartile range (IQR). Two-way repeated measures ANOVA or two-way repeated measures ANOVA on ranks was used, as appropriate, to test safety, feasibility (PRE and POST values), and efficiency.

One-way repeated measures or Friedman repeated measures was used, as appropriate, to test safety and feasibility (POST–PRE differences) and to compare methemoglobin values at different time points.

Paired t-test or Wilcoxon signed rank test was used, as appropriate, to test efficacy. Post-hoc analyses were performed with Bonferroni or Tukey corrections. Statistical significance was defined as p < 0.05. Analysis was performed with SAS software 9.4 (SAS Institute, Inc., Cary, NC, USA) and SigmaPlot v.11.0 (Systat Software Inc, San Jose, CA, USA).

3. Results

3.1. Mathematical Modeling

The PCO₂ of gas/oxygen or distilled water, in a closed system, at increasing concentrations of CO₂ raises linearly, see Figure 2, although the slope is steeper in water relative to gas/oxygen. Instead, if NaOH is added to water, the solution PCO₂ remains close to zero as long as the added CO₂ is lower than the amount of added NaOH. When similar amounts of CO₂ and NaOH are added, almost all CO₂ reacts forming HCO₃⁻ and the solution pH is around 8.220–8.230.

Otherwise, if the added CO₂ is lower than NaOH, carbonic acid dissociates to HCO₃⁻ which, due to the alkaline milieu, further dissociates to CO₃²⁻, thus reducing the concentration of HCO₃⁻. When CO₂ is near half or lower than NaOH, almost all CO₂ forms CO₃²⁻ and the solution pH is above 11. Instead, if the TCO₂ is higher than NaOH, all hydroxide reacts with CO₂ forming HCO₃⁻ and the pH decreases below 8.

Interestingly when the added CO₂ is higher than twice the NaOH, the PCO₂ in the NaOH solution will be higher than the one in a similar gas volume containing the same amount of CO₂.

For example, one liter of gas containing 200 mL (7.86 mmol) of CO₂, the theoretical $\dot{V}$CO₂ of an adult, would have a PCO₂ of 143 mmHg (713 mmHg × 0.2), while 1 L of water would have a higher PCO₂ of 255 mmHg. On the contrary, the same amount of CO₂ could be stored in 1 L of NaOH 10 mmol × min⁻¹ solution with a PCO₂ close to zero.

3.2. Feasibility and Safety Test

No detectable damages to the membrane lung were observed. Moreover, no blood was found in the sweep fluid exiting the oxygenator.

Figure 3 and

Table 1. Efficiency tests results.

Variable		Ventilation	Flow (L × min−1)				p Vent.	p Flow	p Int.
			100	250	500	1000
pH	PRE $	NaOH	7.346 (7.337–7.374)	7.351 (7.333–7.359)	7.356 (7.333–7.363)	7.336 (7.334–7.366)	0.027	0.999	0.020
		O2	7.325 (7.306–7.333) *	7.313 (7.311–7.349) *	7.321 (7.301–7.34) *	7.325 (7.318–7.346)
	POST $	NaOH	7.972 (7.968–8.057) #	7.987 (7.977–8.077) §#	7.964 (7.932–8.040)	7.938 (7.902–8.008)	<0.001	<0.001	<0.001
		O2	7.352 (7.333–7.379) *°§#	7.405 (7.374–7.439) *§#	7.481 (7.435–7.514) *#	7.616 (7.612–7.654) *
	Difference $	NaOH	0.628 (0.597–0.683) #	0.643 (0.624–0.718) §#	0.606 (0.597–0.673)	0.591 (0.565–0.635)	<0.001	<0.001	<0.001
		O2	0.028 (0.011–0.041) *°§#	0.094 (0.063–0.101) *§#	0.145 (0.124–0.186) *#	0.295 (0.268–0.326) *
PCO2 (mmHg)	PRE	NaOH	59.7 (59.2–60.1)	59.5 (59.0–59.7)	59.4 (58.4–60.2)	60.0 (59.2–60.4)	0.909	0.882	0.332
		O2	59.0 (58.4–59.9)	59.0 (58.7–60.5)	60.6 (59.0–61.0)	59.7 (59.5–59.8)
	POST	NaOH	11.2 (11.0–13.0)	10.5 (10.3–11.1)	11.7 (11.3–12.1)	13.1 (12.8–13.1)	<0.001	<0.001	<0.001
		O2	54.6 (53.7–56.2) *°§#	46.2 (45.4–49.7) *§#	40.4 (39.0–41.6) *#	28.2 (26.9–29.1) *
	Difference	NaOH	−48.3 (−48.9–−47.1)	−48.5 (−50.5–−48.3)	−47.5 (−49.0–−46.3)	−46.5 (−47.6–−45.3)	<0.001	<0.001	<0.001
		O2	−4.4 (−6.2–−2.0) *°§#	−13.1 (−13.8–−9.5) *§#	−19.1 (−22.8–−17.4) *#	−31.7 (−32.9–−30.7) *
PO2 (mmHg)	PRE	NaOH	138.0 (136.0–139.0) #	137.0 (137.0–143.0) #	137.5 (136.0–146.0)	141.5 (137.0–153.0)	0.231	0.460	0.002
		O2	144.0 (141.0–162.0)	143.0 (140.0–159.0)	143.5 (138.0–156.0)	143.0 (139.0–154.0)
	POST $	NaOH	125.0 (120.0–130.0)§#	130.5 (128.0–140.0) #	148.5 (142.0–157.0)	161.5 (159.0–169.0)	<0.001	<0.001	0.700
		O2	595.5 (591.0–602.0) *§#	608.5 (603.0–623.0) *#	616.0 (611.0.–6260) *#	648.0 (632.0–654.0) *
	Difference	NaOH	−13.0 (−16.0— −11.0)§#	−6.5 (−9.0–−3.0) §#	9.0 (6.0–11.0) #	18.0 (14.0–21.0)	<0.001	<0.001	0.105
		O2	451.5 (429.0–461.0) *§#	455.5 (445.0–471.0) *§#	462.5 (453.0–477.0) *#	497.5 (487.0–508.0) *
K+ (mEq × L−1)	PRE	NaOH	4.1 (4.0–4.4)	4.1 (4.1–4.5)	4.2 (4.1–4.4)	4.2 (4.1–4.5)	0.127	0.594	0.299
		O2	4.1 (3.9–4.2)	4.0 (4.0–4.2)	4.0 (4.0–4.2)	4.1 (4.0–4.2)
	POST	NaOH	4.1 (4.0–4.3)	4.1 (4.0–4.4)	4.1 (4.0–4.3)	4.1 (4.0–4.4)	0.265	0.709	0.363
		O2	4.1 (3.9–4.2)	4.0 (4.0–4.2)	4.0 (4.0–4.1)	4.0 (3.9–4.2)
	Difference	NaOH	0.0 (0.0–0.1)	0.1 (0.0–0.1)	0.1 (0.1–0.1)	0.1 (0.1–0.1)	0.009	0.337	0.86
		O2	0.0 (0.0–0.0) *	0.0 (0.0–0.0) *	0.0 (0.0–0.0) *	0.0 (0.0–0.1) *
Na+ (mEq × L−1)	PRE	NaOH	143.0 (142.0–144.0)	143.0 (143.0–144.0)	143.5 (142.0–145.0)	143.5 (143.0–144.0)	0.038	0.233	0.973
		O2	139.0 (138.0–143.0) *	139.0 (139.0–143.0) *	139.5 (138.0–144.0) *	139.5 (138.0–145.0) *
	POST	NaOH	141.0 (139.0–142.0)	140.0 (140.0–142.0)	140.5 (140.0–141.0)	141.5 (140.0–142.0)	0.407	0.524	0.096
		O2	139.0 (138.0–143.0)	138.5 (137.0–143.0)	139.0 (138.0–143.0)	138.5 (137.0–143.0)
	Difference	NaOH	−2.0 (−3.0–−2.0)	−3.0 (−3.0–−2.0)	−3.0 (−4.0–−2.0)	−2.0 (−3.0–−2.0)	<0.001	0.215	0.012
		O2	0.0 (0.0–0.0) *#	−1.0 (−1.0–0.0) *	−1.0 (−1.0–−1.0) *	−1.0 (−1.0–−1.0) *
Ca++ (mEq × L−1)	PRE	NaOH	1.3 (1.3–1.4)	1.4 (1.3–1.4)	1.4 (1.3–1.4)	1.4 (1.3–1.4)	0.755	0.854	0.769
		O2	1.4 (1.2–1.4)	1.3 (1.2–1.4)	1.3 (1.2–1.4)	1.3 (1.2–1.4)
	POST	NaOH	1.2 (1.1–1.3)	1.2 (1.1–1.2)	1.2 (1.2–1.3)	1.2 (1.2–1.3)	0.066	0.110	<0.001
		O2	1.4 (1.2–1.4) *§#	1.3 (1.2–1.4) *§#	1.3 (1.2–1.3)	1.3 (1.2–1.3)
	Difference $	NaOH	−0.1 (−0.1–−0.1)	−0.2 (−0.2–−0.1)	−0.1 (−0.1–−0.1)	−0.1 (−0.1–−0.1)	<0.001	0.032	0.002
		O2	0.0 (0.0–0.0) *§#	0.0 (0.0–0.0) *§#	0.0 (0.0–0.0) *	−0.1 (−0.1–−0.1) *
Cl− (mEq × L−1)	PRE	NaOH	111.5 (111.0–113.0)	111.5 (111.0–113.0)	111.5 (111.0–113.0)	111.5 (110.0–113.0)	0.232	0.529	0.529
		O2	111.0 (111.0–112.0)	111.0 (111.0–112.0)	111.0 (110.0–112.0)	111.0 (110.0–112.0)
	POST	NaOH	114.0 (114.0–115.0)	114.5 (114.0–115.0)	114.0 (114.0–115.0)	114.0 (114.0–115.0)	0.002	0.042	0.002
		O2	111.5 (111.0–113.0) *#	111.0 (111.0–113.0) *#	112.0 (111.0— 113.0) *#	112.5 (112.0–114.0) *
	Difference	NaOH	2.5 (2.0–3.0)	3.0 (2.0–3.0)	2.5 (2.0–3.0)	2.5 (2.0–3.0)	0.007	0.002	0.001
		O2	0.5 (0.0–1.0) *#	0.0 (0.0–1.0) *§#	1.0 (1.0–1.0) *#	2.0 (1.0–2.0)
Lac (mEq × L−1)	PRE $	NaOH	1.4 (0.5–2.3)	1.4 (0.5–2.4)	1.3 (0.5–2.3)	1.4 (0.5–2.5)	0.180	0.361	0.614
		O2	1.1 (0.4–2.5)	1.2 (0.4–2.6)	1.1 (0.4–2.6)	1.1 (0.4–2.5)
	POST $	NaOH	1.4 (0.5–2.3)	1.5 (0.5–2.3)	1.4 (0.5–2.4)	1.4 (0.5–2.4)	0.197	0.459	0.850
		O2	1.0 (0.4–2.6)	1.1 (0.4–2.6)	1.2 (0.4–2.6)	1.1 (0.4–2.5)
	Difference	NaOH	0.0 (0.0–0.0)	−0.1 (−0.1–0.1)	0.1 (0.0–0.1)	0.0 (0.0–0.0)	1.000	0.297	0.922
		O2	0.0 (0.0–0.0)	0.0 (−0.1–0.0)	0.0 (0.0–0.1)	0.0 (0.0–0.0)
Hb (g × dL−1)	PRE	NaOH	6.45 (5.50–8.20)	6.80 (5.30–8.20)	6.70 (5.30–8.30)	6.70 (5.20–8.10)	0.643	0.641	0.511
		O2	6.55 (5.50–8.30)	6.60 (5.30–8.20)	6.55 (5.30–8.40)	6.55 (5.40–7.90)
	POST	NaOH	6.60 (5.50–8.30)	6.70 (5.40–8.20)	6.75(5.40–8.30)	6.60 (5.30–8.20)	0.547	0.083	0.893
		O2	6.55 (5.50–8.40)	6.60 (5.30–8.20)	6.60 (5.30–8.50)	6.55 (5.40–7.90)
	Difference	NaOH	0.05 (0.00–0.10)	0 (−0.10–0.00)	0.05 (0.00–0.10)	0.05 (0.00–0.10)	0.025	0.661	0.154
		O2	0.00 (−0.10–0.00) *	0.00 (0.00–0.00) *	0.00 (0.00–0.10) *	0.00 (0.00–0.00) *
HCO3− (mmol × L−1)	PRE	NaOH	30.1 (29.5–32.2)	30.2 (29.3–31.8)	30.3 (29.4–31.7)	29.8 (29.0–31.8)	0.050	0.508	0.165
		O2	28.3 (27.6–29.9) *	28.2 (27.6–29.8) *	28.1 (27.5–30.4) *	28.5 (28.2–30.2) *
	POST	NaOH	25.7 (23.3–29.1)	25.9 (23.2–28.1)	26.4 (23.2–29)	26.3 (23.7–29.7)	0.577	0.043	0.003
		O2	28.0 (27.2–28.6) §#	27.5 (26.7–28.6) #	27.0 (26.3–28.2)	26.6 (26.1–27.2)
	Difference	NaOH	−4.4 (−6.3–−2.4) #	−4.5 (−6.8–−2.4) #	−4.2 (−6.2–−1.3)	−3.9 (−5.7–−1.7)	0.018	0.003	<0.001
		O2	−0.3 (−0.5–−0.2) *§#	−0.8 (−0.9–−0.6) *§#	−1.3 (−1.7–−1) *#	−2.1 (−2.8–−1.4)
plasma TCO2 (mmol × L−1)	PRE	NaOH	31.9 (31.4–34.1)	32.0 (31.1–33.6)	32.1 (31.3–33.5)	31.7 (30.9–33.6)	0.051	0.476	0.194
		O2	30.1 (29.5–31.8)	30.1 (29.4–31.5)	30.0 (29.4–32.3)	30.3 (30.1–32.0)
	POST	NaOH	26.0 (23.6–29.5)	26.3 (23.5–28.4)	26.7 (23.6–29.3)	26.7 (24.1–30.1)	0.258	0.009	<0.001
		O2	29.7 (28.9–30.1) §#	29.1 (28.1–30.0) #	28.3 (27.4–29.5) #	27.5 (26.9–28.1)
	Difference	NaOH	−5.9 (−7.8–−3.9) #	−6 (−8.3–−4.0) #	−5.7 (−7.6–−2.8)	−5.3 (−7.1–−3.3)	0.006	<0.001	<0.001
		O2	−0.5 (−0.6–−0.3) *°§#	−1.1 (−1.3–−0.9) *§#	−1.9 (−2.2–−1.7) *#	−3.0 (−3.8–−2.3)
$\dot{V}$CO2 (mL × min−1)		NaOH	65.3 (43.3–86.7) #	67.0 (44.3–92.2) #	63.5 (31.6–84.5)	59.1 (36.4–79)	0.006	<0.001	<0.001
		O2	5.4 (3.7–6.7) *°§#	12.5 (10.5–14.6) *§#	20.7 (18.6–24.9) *#	33.6 (26.1–42.6)

Abbreviations: PCO₂, partial pressure of carbon dioxide; PO₂, partial pressure of oxygen; Na⁺, sodium; K⁺, potassium; Ca⁺⁺, calcium; Cl⁻, chloride; Lac, Lactate; Hb, hemoglobin; HCO₃-, bicarbonate, TCO₂, total CO₂ content, $\dot{V}$CO₂, amount of carbon dioxide removed by the membrane lung. Data are expressed median (IQR); Differences were computed as POST values–PRE values. p: p values of two-way ANOVA RM or two-way ANOVA RM on ranks (^$) for NaOH vs. O₂ comparison (p Ventilation), Flow effect (p Flow) and interaction (p int.); Post-hoc analysis with Bonferroni or Tukey corrections: * p < 0.05 vs. NaOH; ° p < 0.05 vs. 250 mL/min; § p < 0.05 vs. 500 mL/min; # p < 0.05 vs. 1000 mL × min⁻¹.

report the BGAs of PRE and POST blood. PCO_2PRE was stable throughout the entire test, 59.0 (58.0–60.0) mmHg. Delta PCO₂ across the membrane lung was significantly lower at 10 mmol × L⁻¹ (−32.2 (−38.6–−23.1) mmHg). Otherwise, it showed small increases as NaOH concentration increases (−41.4 (−43.1–−36.8), −47.7 (−49.5–−44), −47.8 (−48.6–−47) and −48.2 (−48.4–−46.6) mmHg at 30, 60, 90, and 100 mmol × L⁻¹ respectively). PCO_2POST was reduced to about 12 mmHg with NaOH concentration ≥ 60 mmol × L⁻¹ (12.0 (11.0–15.0), 11.4 (10.2–12.4) and 12.4 (11.3–13.1) mmHg at 60, 90, and 100 mmol × L⁻¹ respectively), subsequently pH_POST increased up to 7.913 (7.885–7.943) at NaOH concentration equal to 100 mmol × L⁻¹. The lowest $\dot{V}$CO₂ was also recorded at the lowest NaOH concentration 73.9 (54.3–91.8) mL × min⁻¹. PRE blood sodium and potassium concentration were stable (see Supplementary Table S1 for details). POST chloride concentrations were higher than PRE values while sodium concentrations were lower. Moreover, a simultaneous decrease in potassium and calcium POST concentrations was observed. These results are similar to the observations of Langer et al. in couples of measurements of blood entering and leaving the ML in 20 critically ill patients [36]. Methemoglobin values were not different over the time during the experiments (median (IQR) values 1.100 (0.950–2.850) at 15 min, 1.100 (1.050–2.150) at 30 min, 1.100 (1.050–2.400) at 45 min, 1.200 (1.050–2.900) at 60 min; p = 0.606).

As the highest delta PCO₂ was observed when 100 mmol × L⁻¹ NaOH was used, this concentration was employed for the efficiency and efficacy tests.

3.3. Efficiency Test

Blood gas analyses of PRE and POST blood with NaOH and oxygen are reported in Table 1. PCO_2PRE was stable throughout the entire test, 59.6 (58.9–60.4) mmHg. Increasing oxygen flows showed increasing CO₂ removal, both as delta PCO₂ across the membrane lung and $\dot{V}$CO₂, see Figure 4. Conversely, all NaOH flows showed similar CO₂ removal, except for a lower $\dot{V}$CO₂ at 1000 mL × min⁻¹ compared to 100 and 250 mL × min⁻¹ (see Figure 3). When comparing $\dot{V}$CO₂ achieved with liquid and gaseous ventilation, liquid ventilation achieved significantly higher CO₂ removals for 100, 250, and 500 mL × min⁻¹ of flow. On the contrary, while the median value was higher also for 1000 mL × min⁻¹, this difference did not reach statistical significance.

Blood pH_POST increased, according to the PCO₂ reduction, reaching values as high as 7.987 with NaOH at 250 mL × min⁻¹.

PO_2PRE was stable around 141.0 (137.0–147.0) mmHg both during NaOH and oxygen steps while PO_2POST increased up to 470.5 (452.3–507.0) mmHg only during oxygenation use, while it remained unchanged during liquid ventilation.

Blood electrolytes and lactate concentrations were stable throughout the experiment.

3.4. Efficacy Tests

In agreement with the highest $\dot{V}$CO₂ and delta PCO₂, NaOH 250 mL × min⁻¹ was selected as the most performant NaOH flow and compared with 10 L × min⁻¹ of oxygen in the efficacy test.

Table 2. Efficacy tests results.

Variable		Ventilation
		NaOH 250 mL × min−1	O2 10,000 mL × min−1	p
pH	PRE	7.351 (7.333–7.359)	7.328 (7.322–7.355)	0.032
	POST	7.987 (7.977–8.077)	7.966 (7.921–8.013)	0.020
	Difference	0.643 (0.624–0.718)	0.627 (0.599–0.685)	0.094
PCO2 (mmHg)	PRE	59.5 (59–59.7)	60 (59.3–60.5)	0.254
	POST	10.5 (10.3–11.1)	11.5 (10.8–13.9)	0.106
	Difference	−48.5 (−50.5–−48.3)	−48.1 (−49.4–−46.1)	0.522
PO2 (mmHg)	PRE $	137 (137–143)	139 (137–165)	0.625
	POST	130.5 (128–140)	661.5 (649–677)	<0.001
	Difference	−6.5 (−9–−3)	518.5 (509–536)	<0.001
K+ (mEq × L−1)	PRE	4.1 (4.1–4.5)	4.2 (4–4.4)	0.611
	POST $	4.1 (4–4.4)	4.1 (4–4.3)	0.438
	Difference	0.1 (0–0.1)	0.1 (0–0.1)	1.000
Na+ (mEq × L−1)	PRE	143 (143–144)	142 (141–145)	0.516
	POST	140 (140–142)	139.5 (139–143)	1.000
	Difference	−3 (−3–−2)	−2 (−2–−2)	0.102
Ca++ (mEq × L−1)	PRE	1.4 (1.3–1.4)	1.3 (1.3–1.4)	0.927
	POST	1.2 (1.1–1.2)	1.2 (1.1–1.3)	0.413
	Difference	−0.2 (−0.2–−0.1)	−0.1 (−0.1–−0.1)	0.067
Cl− (mEq × L−1)	PRE $	111.5 (111–113)	111.5 (111–112)	0.813
	POST	114.5 (114–115)	115 (114–115)	1.000
	Difference	3 (2–3)	3 (2–3)	0.611
Lac (mEq × L−1)	PRE $	1.4 (0.5–2.4)	1.3 (0.5–2.5)	0.375
	POST	1.5 (0.5–2.3)	1.2 (0.4–2.5)	0.233
	Difference	−0.1 (−0.1–0.1)	0 (−0.1–0)	1.000
Hb (g × dL−1)	PRE	6.80 (5.30–8.20)	6.55 (5.50–8.40)	0.499
	POST	6.70 (5.40–8.20)	6.55 (5.40–7.90)	0.590
	Difference	0.00 (−0.10–0.00)	0.00 (0.00–0.00)	0.363
HCO3− (mmol × L−1)	PRE	30.2 (29.3–31.8)	29.4 (28.7–30.6)	0.205
	POST	25.9 (23.2–28.1)	25.1 (23.1–28.1)	0.652
	Difference	−4.5 (−6.8–−2.4)	−4.6 (−5.7–−1.8)	0.185
plasma TCO2 (mmol × L−1)	PRE $	32 (31.1–33.6)	31.3 (30.6–32.5)	0.313
	POST	26.3 (23.5–28.4)	25.5 (23.5–28.4)	0.695
	Difference	−6 (−8.3–−4)	−6.1 (−7.1–−3.4)	0.191
$\dot{V}$CO2 (mL × min−1)		67 (44.3–92.2)	67.4 (37.8–79.4)	0.191

Abbreviations: PCO₂, partial pressure of carbon dioxide; PO₂, partial pressure of oxygen; Na⁺, sodium; K⁺, potassium; Ca⁺⁺, calcium; Cl⁻, chloride; Lac, Lactate; Hb, hemoglobin; HCO₃⁻, Bicarbonate, TCO₂, total CO₂ content, $\dot{V}$CO₂, amount of carbon dioxide removed by the membrane lung. Data are expressed median (IQR); Differences were computed as POST values–PRE values. p: p values of Paired t-test or Wilcoxon Signed Rank Test (^$) for NaOH (250 mL × min⁻¹) vs. O₂ (10,000 mL × min⁻¹) comparison.

reports blood gas analyses of PRE and POST blood. Both delta PCO₂ and $\dot{V}$CO₂ were similar, suggesting similar extracorporeal CO₂ removal (Figure 4, shadowed boxes).

4. Discussion

This in-vitro study shows that continuous infusion (from 100 to 1000 mL × min⁻¹) of highly concentrated sodium hydroxide solutions into the gas side of conventional polypropylene oxygenators is feasible, despite pH values of the sweeping solution above 12. At low sweep flows, alkaline liquid ventilation showed significantly higher CO₂ removal capacity than conventional gaseous ventilation. However, the maximum CO₂ removal efficiency achieved through liquid ventilation was not superior to the one achieved with 10 L × min⁻¹ of sweep gas flow.

The working hypothesis underlying this study was to exploit the high CO₂ absorbing capacity of NaOH solutions. Indeed, in our experimental context, the concentration of NaOH was always significantly higher than the amount of CO₂ extracted from the ML. The PCO₂ of the alkaline sweep fluid was persistently very close to 0 mmHg, as the added CO₂ was instantly hydrated and dissociated to bicarbonate and carbonate. This allowed to keep the PCO₂ close to zero and thus optimize the transmembrane PCO₂ gradient, favoring the efficiency of extracorporeal CO₂ removal.

Indeed, a solution containing 10 mEq of NaOH could absorb 200 mL of CO₂ while maintaining PCO₂ close to zero. On the contrary, the same amount of CO₂ added to 10 L of gas would result in a PCO₂ around 15 mmHg, therefore reducing the blood-gas CO₂ gradient.

However, the data showed that increasing NaOH flow did not lead to a linear increase in CO₂ removal. Instead, for NaOH flows greater than 250 mL × min⁻ there was an unexpected reduction in CO₂ removal. This reduced efficiency could depend on the density of the sodium hydroxide solution and the mechanics of the membrane lung. Therefore, a clinical application of alkaline liquid ventilation does not seem exploitable using the current technology. The technical complexity and safety profile require further evaluations, although the present tests have recorded no damage to the membrane lung.

Another important difference between gaseous and liquid ventilation needs to be discussed. Although the oxygenation capacity of low-flow devices using conventional gaseous ventilation is limited by the amount of blood reaching the ML, a certain amount of oxygen is added to the blood. On the contrary, the NaOH infusion does not oxygenate the extracorporeal blood, limiting its potential clinical application to patients with isolated hypercapnic respiratory failure, i.e., able to oxygenate properly through their native lungs.

Although devices with higher CO₂ extraction capacity resulted more effective [3,9,10], numerous studies confirm the ability of ECCO₂R to achieve physiological targets. Nevertheless, the clinical application of ECCO₂R is still limited and no conclusive indications have been identified mainly because of safety concerns [37,38]. Indeed, a consistently high rate of complications has been reported, mostly related to hemolysis, bleeding, and thrombosis. In this context, the present study aim was to achieve a highly efficient ECCO₂R technique to ensure a clinical efficacy with limited extracorporeal blood flows, potentially enabling regional anticoagulation [39,40]. The tested technology, which was not developed for alkaline liquid ventilation, did not meet such expectations. However, we cannot exclude that a dedicated device could achieve more satisfying results.

This study presents several limitations. First, we could not perform any gas analysis of the sodium hydroxide solution. The CO₂ extraction capacity was estimated both as differences in PCO₂ and TCO₂ between the blood samples upstream and downstream of the artificial lung [41]. $\dot{V}$CO₂ showed higher variability than PCO₂, as shown in Figure 4, possibly due to the baseline different blood composition. Indeed, we can speculate that this phenomenon may be explained by the wide range in hemoglobin concentrations (see Table 1), which affects the ML $\dot{V}$CO₂ [42]. Secondly, the alkaline liquid ventilation was tested only in vitro and for a limited time consequently we cannot exclude different effects and safety issues in vivo scenarios. Thirdly, we only tested one type of polypropylene membrane lung. Further tests with different devices may be required.

5. Conclusions

This in-vitro study showed that ECCO₂R through alkaline liquid ventilation of the ML is feasible and safe. The CO₂ removal efficiency of alkaline liquid ventilation was higher than conventional gaseous ML ventilation only for low sweep flows. Indeed, at high sweep gas flow, the CO₂ removal efficiency was comparable between the two techniques.

The development of a dedicated device may be necessary to exploit the potential of this technology. Further studies will be required before any possible clinical application.