This randomized experimental trial was conducted at the Hedenstierna Laboratory (Uppsala University, Sweden) and is reported in accordance with the Arrive 2.0 Guidelines [14] and a checklist is included as Additional file 1. The study was approved by the Animal Ethics Board in Uppsala, Sweden (Dnr. 5.8.18-05377/2021).

Healthy pigs aged 3–5 months of the Norwegian landrace/Yorkshire/Hampshire mixed breed, were included in the experiment. The pigs were randomized into two groups at a 1:1 ratio using the Research Randomizer online software [15] prior to the start of the experiments. One group of pigs received continuous chest compressions and continuous ventilations (10/min) asynchronous with the chest compressions (CCC group). In this group, a timer was used to give a signal every 6th second to indicate when the ventilations were to be given (corresponding to 10 ventilations/min). The second group received 30 chest compressions followed by a 3.2 s pause, during which 2 ventilations were given (corresponding to 6 ventilations/min) (30:2 group). The chest compressions in both groups were given with a frequency of 102/min using a LUCAS 3 mechanical chest compression device (Jolife AB/Stryker, Lund, Sweden).

A Laerdal Bag II (maximum measured volume: 1650 ml) (Laerdal medical, Stavanger, Norway) was used for the manual ventilation. The same person ventilated all of the pigs in the study. A portable respiratory monitor based on pneumotachography (Fluxmed GrH®, MBMed, Buenos Aires, Argentina) was used to both record the ventilation data and guide the ventilations with the aim of each ventilation being as close to the intended tidal volume of 8 ml/kg of weight as possible.

Due to the nature of this study, the researchers who performed the experiment were not blinded to the randomization.

ProtocolPreparation

The pigs fasted for 12 h before the experiment, with free access to water. Upon the arrival to the laboratory, they were weighed, and their status checked (signs of stress, wounds or dehydration). They were then anesthetized with a subcutaneous injection of tiletamine (6 mg/kg) and zolazepam (2.2 mg/kg). After 5 to 10 min, they were placed on an operating table where an ear vein was cannulated to administer a bolus of fentanyl (10–20 mg/kg). The pigs were tracheotomized and an endotracheal tube inserted. One neck artery and one femoral artery were cannulated (one was connected to the monitor through a transducer for pressure measurements and the other was used to withdraw blood gas samples). A jugular vein was also cannulated and connected to the monitor to facilitate central venous pressure measurements. A small incision was performed to expose the urinary bladder for insertion of a urinary catheter.

Anesthesia was maintained throughout the experiment with a continuous infusion of ketamine (30 mg kg−1 h−1), midazolam (0.1 mg kg−1 h−1) and fentanyl (3.75 mg kg−1 h−1). After checking that the anesthesia depth was sufficient to prevent responses to painful stimuli, a continuous infusion of rocuronium (following a bolus dose of 0.6 mg/kg) was started. During the one-hour preparation time an infusion of Ringer’s acetate was administered with 30 ml kg−1 h−1 to prevent dehydration. Mechanical ventilation was performed (Servo I, Maquet, Stockholm, Sweden) in volume-controlled mode with Vt 6-8 ml/kg, respiratory rate 25, inspiratory:expiratory ratio 1:2, FiO2 0.5 and PEEP 5 cmH20. At the end of the surgical preparation, the pigs were left to rest for 30 min.

Experiment

Before the induction of cardiac arrest, the mechanical chest compression device was placed on the pigs’ thorax. To ensure adhesion of the suction cup, the edge of the cup was glued to the chest using handball glue. Once the device was positioned, the pig’s hooves were fixated to the table using tape and sandbags were placed between the pig’s flanks and the legs of the device, to prevent any unwanted movement. A baseline arterial blood gas (ABG) was taken.

Ventricular fibrillation (VF) was induced by passing an alternating current of 200 mA for 5 s through the heart of the pig. VF was verified on the electrocardiogram and by the loss of arterial blood pressure.

The pigs were then left untreated for 3 min. During these 3 min the ventilator was disconnected, the portable respiratory monitor and ventilation bag were connected to the endotracheal tube and oxygen connected to the bag with a flow of 12 l/min.

CPR was started according to the randomization. Every 5 min, an ABG sample was taken. Arterial pressure, central venous pressure and ventilation parameters were monitored and recorded continuously during the experiment. After 20 min of CPR, the pigs were euthanized with an injection of potassium chloride (KCl). The chest compressions were continued for a short time after the administration of KCl, to facilitate proper perfusion of the heart. The timeline of the protocol can be found in Fig. 1.

Fig. 1

Timeline of the protocol. CPR cardiopulmonary resuscitation, VF ventricular fibrillation

Data curationVentilation data

The portable respiratory monitor was used for monitoring and recording of ventilation data at a rate of 256 Hz. The device measured time, gas flow, respiratory volume, airway pressure, and exhaled CO2. An R-script (Additional file 2) was written to extract the peak values for each of these parameters as well as the duration of the inspiration for each individual ventilation resulting in the following parameters: peak inspiratory flow (PIF), inspiratory tidal volume (Vti), peak inspiratory pressure (PIP), duration of inspiration (Ti) and peak expiratory CO2 (PECO2). PECO2 was used as a proxy for EtCO2.

A mean value was then calculated for each ventilation parameter/minute of the experiment (20 in total/parameter) and used for the analysis.

SpO2 was monitored before the induction of cardiac arrest and was included in the baseline values.

Hemodynamic data

Arterial pressure and central venous pressure were monitored continuously throughout the experiment and recorded at 125 Hz.

Data from 5 min before the induction of cardiac arrest were used to calculate Baseline values. Heart rate was included only as a baseline measurement as the chest compression device performed chest compressions with a standard frequency of 102 per minute throughout the experiment.

For each compression/decompression cycle, two time points were identified, maximum compression and maximum decompression. At the time of maximum compression, maximum arterial pressure (maxAP) and maximum central venous pressure (maxCVP) were recorded. At the time of maximum decompression, minimum arterial pressure (minAP) and minimum central venous pressure (minCVP) were recorded. In order to identify these values, a MatLab script was written and applied to the collected data (Additional file 3).

Coronary perfusion pressure (CPP) was calculated at each compression/decompression cycle as the difference between arterial and central venous pressure at the end of decompression [16]. A mean value per minute was then calculated for each hemodynamic parameter and used for the analysis.

Post-mortem analysisNecroscopic examination

After the experiment all animals underwent a necroscopic examination and the incidence of injuries were recorded. A more in-depth assessment of the lungs was also performed, with the visual grading of atelectasis, signs of hyperinflation and hemorrhages (graded 0 to 3). The examination and grading system to assess the injuries are described in detail in Additional file 4.

Histopathological analysis

After the necroscopic examination, lung tissue samples were taken from 5 parts of both the left and right lung and put in formaldehyde. The sampled locations were: paracardiac region, upper ventral lobe, upper dorsal lobe, lower ventral lobe and lower dorsal lobe (10 samples per subject). The samples were then shipped to the National Veterinary Institute (SVA—Uppsala, Sweden) where a pathologist, blinded to the protocol, examined them with microscopy and evaluated the following features:

1.

Atelectasis (percentage of atelectasis tissue over the whole tissue analyzed in the sample, and atelectasis pattern—patchy or homogeneous).

2.

Hyperinflation (percentage of hyperinflated tissue over the whole tissue analyzed in the sample).

3.

Edema (assessed using a grading system from 0 to 5).

4.

Microscopic hemorrhages (only the presence or absence of the feature was reported, as a binary variable).

A full description of the histopathological analysis can be found in Additional file 5.

Wet–dry ratio

The degree of edema in the lung tissue was also measured using wet–dry ratio. Samples were taken from the same locations as for the histopathological examination. They were weighed fresh, dried in an oven at 37° for a week, and then re-weighed.

Study variables

Peak inspiratory pressure (PIP).

Max and min arterial pressure (maxAP, minAP).

Max and min central venous pressure (maxCVP, minCVP).

Coronary perfusion pressure (CPP).

Inspiratory tidal volume (Vti).

Duration of inspiration (Ti).

Minute volume (MV).

Peak expiratory CO2 (PECO2) (as a measurement of both hemodynamics and ventilation).

Peak inspiratory flow (PIF).

Arterial blood gases (ABG).

Changes over time in hemodynamics, PECO2 and ABG during the experiment.

Edema assessed by wet–dry ratio.

Histopathological assessment of the lungs (atelectasis, edema and hyperinflation).

Statistical analysis

Baseline values and the values from the post mortem analyses were based on one value per pig and analyzed using the Mann–Whitney U-test for continuous variables and Fishers’ exact test for categorical variables. Descriptive statistics were presented as n for categorical variables and median and IQR for continuous variables. The multiple comparisons of lung tissues were adjusted using the Bonferroni correction.

Differences in the mean values throughout the experiment and changes over time were analyzed using linear mixed models with time as a repeated measure and CPR mode, time, and CPR mode-time interaction as fixed factors. The covariance structure for the models was chosen based on analysis of model fit using Akaike’s information criterion (AIC). First order ante-dependence was found to minimize AIC in a plurality of the models and was therefore used in all the mixed models. Comparison of model fits using various covariance structures may be found in Additional file 6. Mean values, confidence intervals, and p-values were derived from the models fixed effect estimates.

Correlations were analyzed with Pearson’s test with one mean value per parameter and subject being used.

A p-value of < 0.05 was considered significant.

Statistical analyses performed using SPSS 28.0 (IBM©).