The VLP2000E is a pneumatically powered, time-cycled, and pressure-limited mechanical ventilator weighing 150 grams. The device works automatically when connected to a portable oxygen tank, with easy adjustment of the respiratory rate. In addition, ventilation can also be manually triggered by pushing a button. It has mechanisms that allow spontaneous breathing, as well as a special valve that limits PIP even during simultaneous ventilation with chest compressions.

Female Landrace pigs weighing approximately 33 kg were fasted for 12 hours with free access to water prior to the procedure. After intramuscular anesthesia (5 mg/kg of ketamine hydrochloride [Ketalar 50 mg/mL, Pfizer]; 0.5 mg/kg of midazolam [Dormonid 5 mg/mL, Roche]), 12.5 mg/kg of thiopental (Thiopentax 20 mg/mL, Cristália) was infused through an ear vein, and the animal was intubated with a number seven endotracheal tube with cuff, which was inflated to avoid leaks. Continuous intravenous anesthesia was maintained with midazolam 1.5 mg/kg/h and 0.015 mg/kg/h of fentanyl citrate (Fentanyl 0.05 mg/mL, Janssen) plus 4-mL unrestricted boluses, and thiopental 0.6-6 mg/kg/h.

A Dräger Evita XL conventional mechanical ventilator (Drägerwerk AG & Co. KGaA, Lübeck, Germany) was used only during preparation, which was set to deliver intermittent positive pressure, tidal volume of 10 mL/kg, positive end-expiratory pressure (PEEP) of 5 cm H₂O, and fraction of inspired oxygen (F_IO2) of 50%; the respiratory rate was adjusted to maintain the end-tidal carbon dioxide (ETCO₂) between 40-45 mm Hg.

The animals’ jugular veins and femoral arteries were punctured and cannulated with 8 French (F) and 6 F hemostatic introducers, respectively. The right jugular vein was used to infuse general anesthesia and all medications. A Swan-Ganz catheter of 7.5 F was placed with continuous cardiac output in the pulmonary artery via left jugular vein, and a pigtail catheter of 6 F was placed in the aortic root via right femoral artery.

The following equipment items were used: Dräger Infinity Delta XL multiparameter monitor (Dräger Medical Systems Inc., Telford, PA, USA); NICO ETCO₂monitor (Koninklijke Philips Electronics NV, Amsterdam, Netherlands); Vigilance II cardiac output monitor (Edwards Lifesciences LLC, Irvine, CA, USA); MVA1000 airway monitor (Neurony Ltda., Vitória, ES, Brazil); Biopac MP100 data acquisition system with AcqKnowledge software (Biopac Systems Inc., Goleta, CA, USA); and Radiometer ABL735 blood gas analyzer (Radiometer Medical ApS, Brønshøj, Denmark).

The Biopac system recorded the following parameters at 250 samples per second: electrocardiogram (Biopac surface electrodes); aortic pressure (Biopac pressure transducer connected to pigtail catheter); right atrial pressure (Biopac transducer connected to Swan-Ganz atrial port); coronary perfusion pressure (CPP) (the difference between aortic and right atrial pressures); ETCO₂(NICO analogue output connected to Biopac); cardiac output (Vigilance analogue output); and airway pressure and airway flow (MVA1000 analogue outputs). Tidal volume (integral of the airway flow curve at inspiration) and a simplified dynamic lung compliance (volume-pressure ratio at inspiration) were calculated. Central temperature was also monitored.

Blood samples were collected from left femoral artery (or from aorta during CPR) 30 minutes before the cardiac arrest (baseline), in CPR cycles 2, 4, 7, 9, 12, and 14, and after the return of spontaneous circulation (ROSC) at 10, 30, 60, 90, and 120 minutes. Blood gas and respiratory, and hemodynamic parameters (except for cardiac output, which is not usually measurable in CPR) were measured at these cycles and time points.

In all cycles and time points, three consecutive breaths were selected using AcqKnowledge software to measure each parameter and obtain the final mean of the respiratory and hemodynamic parameters. The following intervals were defined for measurement: the total respiratory cycle, the interval between two QRS complexes or two chest compressions, and the inspiration interval, providing the respiratory and heart rates, the chest compression rate, the PIP, the peak inspiratory flow (PIF), the tidal volume and the dynamic lung compliance. In the total respiratory cycle interval, ETCO₂, aortic pressure, right atrial pressure, CPP, and cardiac output were measured.

Ventricular fibrillation was induced by applying direct current from a 9-V battery to the right ventricle for one second via a transvenous pacemaker electrode placed temporarily via right jugular vein. After seven minutes of inactivity to simulate the response time of emergency medical services, 2-minute CPR cycles (100 manual chest compressions per minute with approximately 5 cm deep and 10 breaths per minute) were initiated. Ten pigs were ventilated using an adult bag-valve with reservoir and oxygen influx at 10 L/min, and another 10 pigs were ventilated using the VLP2000E with F_IO2of 100%. In both groups, PEEP was not applied, and oxygen was provided by a portable cylinder. Guided by a metronome, to avoid interpersonal bias, the same team member compressed the chest at a rate of 100 per minute; another member delivered 10 breaths per minute with bag-valve, delivering each breath in approximately two seconds following four beats of the metronome, and delivering the next breath on the 10^thbeat. The VLP2000E was adjusted to deliver 10 breaths per minute. After the first CPR cycle, defibrillation was attempted with 150 J, and CPR was resumed immediately. The cardiac rhythm was checked after each cycle, followed by another attempt with 150 J or CPR resumption. In cycles 2, 4, 7, 9, 12, and 14, adrenaline (Epinephrine 1 mg/mL, Cristália) 0.02 mg/kg was administered. In cycles 3 and 5, amiodarone (Ancoron 50 mg/mL, Libbs) 5 mg/kg and 2.5 mg/kg, respectively, were administered in cases of persistent ventricular fibrillation. CPR was continued for up to 30 minutes or until ROSC, and in cases of ROSC, all animals were placed on the VLP2000E for two hours before euthanasia with anesthetics plus 10 mL of potassium chloride. Sustained ROSC was indicated by an adequate cardiac rhythm and systolic pressure above 50 mm Hg for more than 10 minutes, and if spontaneous cardiac arrest occurred thereafter, the experiment was ended; however, if this occurred within less than 10 minutes, resuscitation was restarted until completing 30 minutes of CPR or until ROSC was achieved. Immediately after each ROSC episode in both groups, five vigorous consecutive breaths were delivered with bag-valve to expand the lungs equally in both groups.

Adopting the assumption that resuscitation would not be impaired by mechanical ventilation, the sample size was calculated assuming a ROSC rate of 83%. Otherwise, if only 25% ROSC was achieved in the mechanical ventilation group, then 10 animals per group would be required for 80% statistical power and 5% two-tailed alpha.

The chi-square test or Fisher’s exact test, two-way ANOVA (time point, group), and Bonferroni multiple comparisons were performed using BioStat statistical software (AnalystSoft, Inc.). Graphs were generated with the Number Cruncher Statistical System (NCSS, LLC), and data are presented as mean (standard deviation) [mean (SD)]. Differences were considered significant at p < 0.05.

Similar (p = 1.0000) ROSC rates were achieved with bag-valve ventilation (60%; three pigs after two CPR cycles, three pigs after three cycles) and VLP2000E ventilation (50%; two pigs after two CPR cycles, three pigs after three cycles). After 30 minutes of observation, three animals in the bag-valve group presented with spontaneous cardiac arrest. At 2 hours of observation, three pigs from the bag-valve group and five pigs from the VLP2000E group remained alive (p = 0.6499).

Some parameters did not differ between the bag-valve and VLP2000E groups in all CPR cycles, including the chest compression rate [98 (2) vs. 99 (2) cpm, p = 0.8820], respiratory rate [11 (1) vs. 11 (1) rpm, p = 0.4477], and CPP, for instance in the cycle 1 [30 (17) vs. 35 (16) mm Hg, p = 0.6207], respectively. Blood gas parameters did not differ between the groups in most CPR cycles, except for initial relative hypercapnia in the VLP2000E group, as shown in Table 2 .

Table 2

CPR	GROUP	PaCO2	p	PaO2	p	SaO2	p
		mm Hg		mm Hg		%
Cycle 2	Bag-valve	27 (8)	0.0056	149 (63)	0.1040	97 (3)	0.0912
	VLP2000E	52 (13)		103 (90)		86 (13)
Cycle 4	Bag-valve	35 (12)	0.0337	109 (99)	0.3025	79 (24)	0.3232
	VLP2000E	64 (19)		65 (20)		69 (22)
Cycle 7	Bag-valve	29 (2)	0.0111	83 (24)	0.6193	90 (7)	0.0105
	VLP2000E	72 (20)		56 (13)		58 (21)
Cycle 9	Bag-valve	94 (92)	0.6098	41 (44)	0.8552	45 (59)	0.8714
	VLP2000E	85 (21)		51 (17)		47 (28)
Cycle 12	Bag-valve	97 (107)	0.9224	40 (51)	0.8654	47 (59)	0.8111
	VLP2000E	96 (22)		49 (18)		44 (28)
Cycle 14	Bag-valve	105 (107)	0.9594	35 (35)	0.8477	38 (50)	0.9171
	VLP2000E	104 (20)		45 (16)		37 (24)

Values presented as mean (SD); a: arterial blood; P_CO2: partial pressure of carbon dioxide; P_O2: partial pressure of oxygen; S_O2: oxygen saturation.

Respiratory parameters were significantly affected by the ventilation device, as shown on a cycle-by-cycle basis in Figure 2 . In 14 CPR cycles, the mean of all measurements of each parameter showed significant differences (all p < 0.0001) between bag-valve group and VLP2000E group in PIP [52 (9) vs. 39 (5) cm H₂O], tidal volume [0.635 (0.172) vs. 0.306 (0.129) L], ETCO₂[14 (8) vs. 27 (9) mm Hg], and PIF [0.878 (0.234) vs. 0.533 (0.105) L/s], in addition to inspiratory time [1.96 (0.30) vs. 1.79 (0.18) s, p = 0.0154), respectively.

Figure 2

After ROSC, all animals were placed on the VLP2000E; the parameter values at 10 minutes are shown in Table 3 . At this time point, a significant decrease of dynamic compliance was observed in the bag-valve group versus the VLP2000E group [0.016 (0.006) vs. 0.022 (0.004) L/cm H₂O, p = 0.0262]. Despite the use of the same ventilator and similar respiratory rates, flow levels, and airway pressures after ROSC, the animals in the VLP2000E group had higher tidal volume than the animals ventilated with bag-valve during CPR, as shown in Figure 3 . Dynamic lung compliance was ≥ 0.025 L/cm H₂O prior to cardiac arrest and, following ROSC, it decreased significantly after CPR with bag-valve ventilation, but not after CPR with VLP2000E ventilation ( Figure 4 ). A summary of the study is shown in Central Illustration .

Table 3

PARAMETER	BAG-VALVE	VLP2000E	p
	6 pigs	5 pigs
PaCO2(mm Hg)	37 (15)	41 (8)	0.7307
PaO2(mm Hg)	190 (119)	265 (106)	0.0529
SaO2(%)	94 (10)	98 (4)	0.6720
ETCO2(mm Hg)	26 (14)	31 (9)	0.2004
Respiratory rate (rpm)	19 (2)	19 (2)	0.9719
Heart rate (bpm)	89 (45)	128 (62)	0.0022
Cardiac output (L/min)	1.9 (0.3)	2.0 (0.3)	0.7483
Aortic pressure (mm Hg)	57 (43)	60 (24)	0.8819
Right atrial pressure (mm Hg)	17 (3)	11 (4)	0.0056
PIP (cm H2O)	21 (4)	19 (3)	0.7172
PIF (L/s)	0.628 (0.138)	0.729 (0.066)	0.2498
Tidal volume (L)	0.310 (0.096)	0.413 (0.091)	0.2387
DCompliance (L/cm H20)	0.016 (0.006)	0.022 (0.004)	0.0262

Values presented as mean (SD); a: arterial blood; P_CO2: partial pressure of carbon dioxide; P_O2: partial pressure of oxygen; S_O2: oxygen saturation; ETCO₂: end-tidal carbon dioxide; PIP: peak inspiratory pressure; PIF: peak inspiratory flow; DCompliance: dynamic compliance.

Figure 3

Figure 4

Central Illustration A summary of the study.

The frequent problem of hyperventilation in CPR can be avoided with the use of the VLP2000E. It also provides automatic ventilation, which allows a reduction in the number of rescuers, and the use of a high-efficiency particulate air (HEPA) filter between the endotracheal tube (or face mask) and the special valve reduces the risk of aerosolization of viral particles.²³ Therefore, this ventilator may be appropriate for the safety of rescuers when treating patients with cardiac arrest with suspected or confirmed COVID-19.²³ These qualities may render the VLP2000E a noteworthy option as ventilation device, especially in the field, transport units, acute diseases, and CPR.

The animals were healthy prior to the experiment, and this model of CPR does not simulate cases of patients with chronic obstructive pulmonary disease or asthma. Lung biopsy or necropsy specimens were not analyzed, nor were the lungs X-rayed; such measures may contribute to the understanding of potential causes of decreased lung compliance. In addition, the experiment was not planned for late observation, which could provide more information on the progression of dynamic compliance and on cardiac arrest outcomes.