Ventilator settings and stimulation pattern identification

Until monitoring and phrenic nerve stimulation were established, the pigs were ventilated in a lung-protective pressure-controlled MV mode (tidal volume: 4–8 ml/kg BW, Positive end-expiratory pressure (PEEP) 5 cmH₂O, driving pressure < 15 cmH₂O; Inspiration:Expiration ratio 1:1 to 1:2; respirator rate: 20–26, end-tidal CO₂ 35–45 mmHg, Fraction of inspired oxygen (FiO₂): 0.3) by the ICU-ventilator. The stimulation electrode placement was followed by iterative testing of different PNS stimulation patterns to identify the most effective ones as described in our previous paper. In brief, the stimulation pattern identification was primarily based on the achievable tidal volume. Secondly, when evaluating stimulation patterns with the same tidal volume, a lower pulse length and pulse frequency was preferred, as this results in less electrical energy output to surrounding tissue and the phrenic nerve. In addition, a longer voltage rise time was positively rated, as it was associated with a smoother inspiratory muscle contraction²¹.

The first pig (Pig1) was primarily used for the initial establishment of the ultrasound-guided PNS breathing and systematic stimulation pattern testing. From the second pig (Pig2) onwards, six hours of the 24 h experimental period were used for the breath-by-breath comparison with MV.

During PNS, the ICU-ventilator was set to continuous positive airway pressure (CPAP) mode with a PEEP of 5 cmH₂O to avoid atelectasis. However, no pressure support was administered by the ICU-ventilator to allow quantification of PNS-induced diaphragmatic contraction via resulting tidal volumes. After completion of the exploratory stimulation pattern identification, a breath-by-breath alternation between MV and PNS was performed to allow direct comparison of the two ventilation modes. The target tidal volume for phrenic nerve stimulated breathing was set and modulated between 4 and 8 ml/kg BW, to meet the criteria for lung-protective ventilation²³. This tidal volume is primarily administered in patients suffering from acute respiratory distress syndrome (ARDS)²⁴.

The sonographic Brightness- (B-mode) and Motion-mode (M-mode) were used to visualize and evaluate the sufficiency of the diaphragmatic contraction by phrenic nerve stimulation. For quantification of diaphragmatic thickening, the linear ultrasound probe was placed with craniocaudal orientation on the anterior-axillary line in the zone of diaphragmatic apposition to the thoracic wall and M-mode was recorded (position of the sonographic probe is illustrated in Fig. 1a)²⁵. Subsequently, the calculation of the ratio between end-inspiratory and end-expiratory diaphragm thickness (d_I/d_E-ratio) was performed bilaterally from the third pig onwards at two different time points (at the beginning of the experiment and after about twelve hours). To reduce observer bias, quantification was carried out by two authors (MMD, DZ), blind to each other.

Electrical impedance tomography (EIT, PulmoVista 500, Draeger, Luebeck, Germany) was used in two pigs (Pig5, Pig6) to study the regional air distribution in the lungs during ventilation. For this purpose, the EIT belt was placed, so that caudal parts of the lungs were visualized, as dorso-caudal regional ventilation was expected to be reduced, at least during mechanical ventilation²⁶. For the qualitative analysis of EIT impedance signals, data was analyzed using the manufacturer’s software (Draeger EIT Data Analysis Tool 6.1, Draeger, Luebeck, Germany). To quantify the regional distribution of tidal volumes in the caudal lung areas, the geometrical Center of Ventilation (CoV) in EIT was calculated using a custom-built MATLAB script. The CoV is described by x- and y-coordinate pairs each between 0 and 1. For this study, the y-value was relevant. It describes the ventro-dorsal orientation, where larger values represent a more dorsally located CoV²⁷. The CoV was calculated for 300 breaths of PNS as well as MV separately in a 1:1 breath-by-breath comparison.

To describe the relationship between pulmonary ventilation pressure and tidal volume during inspiration and expiration, corresponding curves were calculated. Therefore, from the second pig (Pig2) onwards for each animal automatically recorded ventilator data for 100 stimulated and 100 mechanical breaths were analyzed, and the mean pressure–volume curves were visualized using a custom-built MATLAB script. To determine the effective ventilation pressure applied to the lungs, also called intratracheal pressure (Ptrach), the corresponding airway pressure (Paw) was corrected for artificial endotracheal tube and breathing filter-related resistances as first described by Guttmann et al^28,29. Mean intratracheal pressure values for in- and expiration were calculated for 25%, 50% and 75% of the average PNS-induced tidal volume for statistical analysis.