EDITOR’S PERSPECTIVE
What We Already Know about This Topic

Positive end-expiratory pressure protects against ventilation-induced lung injury by improving homogeneity of ventilation, but positive end-expiratory pressure contributes to the mechanical power required to ventilate the lung

What This Article Tells Us That Is New

This in vivo study (36 pigs mechanically ventilated in the prone position) suggests that low levels of positive end-expiratory pressure reduce injury associated with atelectasis, and above a threshold level of power, positive end-expiratory pressure causes lung injury and adverse hemodynamics

Background

Positive end-expiratory pressure is usually considered protective against ventilation-induced lung injury by reducing atelectrauma and improving lung homogeneity. However, positive end-expiratory pressure, together with tidal volume, gas flow, and respiratory rate, contributes to the mechanical power required to ventilate the lung. This study aimed at investigating the effects of increasing mechanical power by selectively modifying its positive end-expiratory pressure component.

Methods

Thirty-six healthy piglets (23.3 ± 2.3 kg) were ventilated prone for 50 h at 30 breaths/min and with a tidal volume equal to functional residual capacity. Positive end-expiratory pressure levels (0, 4, 7, 11, 14, and 18 cm H2O) were applied to six groups of six animals. Respiratory, gas exchange, and hemodynamic variables were recorded every 6 h. Lung weight and wet-to-dry ratio were measured, and histologic samples were collected.

Results

Lung mechanical power was similar at 0 (8.8 ± 3.8 J/min), 4 (8.9 ± 4.4 J/min), and 7 (9.6 ± 4.3 J/min) cm H2O positive end-expiratory pressure, and it linearly increased thereafter from 15.5 ± 3.6 J/min (positive end-expiratory pressure, 11 cm H2O) to 18.7 ± 6 J/min (positive end-expiratory pressure, 14 cm H2O) and 22 ± 6.1 J/min (positive end-expiratory pressure, 18 cm H2O). Lung elastances, vascular congestion, atelectasis, inflammation, and septal rupture decreased from zero end-expiratory pressure to 4 to 7 cm H2O (P < 0.0001) and increased progressively at higher positive end-expiratory pressure. At these higher positive end-expiratory pressure levels, striking hemodynamic impairment and death manifested (mortality 0% at positive end-expiratory pressure 0 to 11 cm H2O, 33% at 14 cm H2O, and 50% at 18 cm H2O positive end-expiratory pressure). From zero end-expiratory pressure to 18 cm H2O, mean pulmonary arterial pressure (from 19.7 ± 5.3 to 32.2 ± 9.2 mmHg), fluid administration (from 537 ± 403 to 2043 ± 930 ml), and noradrenaline infusion (0.04 ± 0.09 to 0.34 ± 0.31 μg · kg−1 · min−1) progressively increased (P < 0.0001). Lung weight and lung wet-to-dry ratios were not significantly different across the groups. The lung mechanical power level that best discriminated between more versus less severe damage was 13 ± 1 J/min.

Conclusions

Less than 7 cm H2O positive end-expiratory pressure reduced atelectrauma encountered at zero end-expiratory pressure. Above a defined power threshold, sustained positive end-expiratory pressure contributed to potentially lethal lung damage and hemodynamic impairment.

Because expenditure of energy is required to inflict damage, we recently proposed mechanical power (the intensity of energy delivery to the respiratory system) as a unifying concept that includes all primary ventilator settings shown experimentally to influence ventilator-induced lung injury.1  The equation for mechanical power is the product of ventilating frequency and the inflation energy of the tidal cycle. The latter consists of three components: (1) the power required to overcome tissue and airways resistance during gas movement (flow-resistive work); (2) the power required to inflate the lung and chest wall from their shared initial position (tidal volume–associated work); and (3) the power required to overcome positive end-expiratory pressure (PEEP)–related recoil of the lung and respiratory system. In simplified form, the airway pressure developed during inflation can be written as follows:

Multiplying each of these components by the change in volume (Vt), we obtain the mechanical energy per breath (Energytot).

In turn, this energy per cycle, when multiplied by the respiratory rate, yields the mechanical power applied per minute to the respiratory system.

In a previous long-term experiment on healthy piglets, we tested the following elements of the mechanical power: tidal volume (strain),2  respiratory rate,3  and flow.4  We found that the tidal volume was predictably lethal when it reached the total lung capacity at 15 breaths/min,2  whereas no major damage was observed at 6 breaths/min.3  Furthermore, we found an association between the flow rate and the extent of ventilator-induced lung injury.4  Therefore, in that series of experiments, all performed at zero PEEP, we confirmed the roles of the tidal volume, respiratory rate, and flow components of total power as probable contributors to ventilator-induced lung injury.5 

Notably, at 15 breaths/min, the otherwise lethal strain did not induce any marked damage if 75% of the maximal distending volume (i.e., the sum of tidal volume and PEEP volume) was due to PEEP.6  In those experiments, therefore, PEEP appeared to be protective, as shown by Webb and Tierney in the early 1970s.7  It remains unclear, however, whether PEEP is protective per se or whether its putative benefit is due to the associated reductions in tidal volume, driving pressure, and atelectrauma. Actually, because PEEP is a key element of the power equation, it theoretically has lung-damaging potential. To begin inflation, the lung requires an energy input greater than the potential energy stored in the system by PEEP at end exhalation. Recruitment diminishes and distention increases as airway pressure rises. Therefore, although its mechanical effects on atelectrauma may be, on balance, “lung-protective” over its lower range, rising PEEP is unquestionably a component of mechanical power and, as such, should favor ventilator-induced lung injury by increasing lung stress and strain.

In this study, we aimed to investigate the purely mechanical role of PEEP on healthy lungs under the unifying framework of the mechanical power hypothesis. Indeed, we wanted to test whether PEEP per se protects or contributes to ventilator-induced lung injury and, if so, over what range and to what extent.

Thirty-six domestic piglets (~4 months of age; body weight, 23.3 ± 2.3 kg) were handled according to the European Union guidelines 2010/63 with the approval of the local authorities. The experiments were performed under general anesthesia with sufentanil (2 to 3 μg · kg−1 · h−1), propofol (6 to 9 mg · kg−1 · h−1), and midazolam (1.2 to 1.5 mg · kg−1 · h−1). The animals were studied in the prone position and were instrumented with endotracheal tube, esophageal balloon, central venous, pulmonary artery, femoral artery, and urinary catheters. Infusions of Sterofundin 1/1 (B. Braun Melsungen, Germany) of 2 to 3 ml · kg−1 · h−1 were maintained during the whole experiment. When necessary, colloids (Gelafundin 4%, Braun, Germany) and norepinephrine were administered to maintain a mean arterial pressure above 60 mmHg (see Supplemental Digital Content for details, https://links.lww.com/ALN/B790).

Experimental Design

The sample size was based on our experience from previous experiments. All animals were ventilated at a constant respi ratory rate of 30 breaths/min, fractional inspired oxygen tension (Fio2) of 0.4, and ratio between inspiratory and expiratory time of 1:2 with a tidal volume equal to the functional residual capacity (FRC) measured at baseline at zero PEEP, which corresponds to a dynamic strain of tidal volume/FRC = 1 (mean tidal volume = 14.9 ± 2.5 ml/kg). Six different levels of PEEP were applied, one PEEP level to each of six groups of randomly assigned piglets. The planned experimental duration was 50 h, during which all ventilation and PEEP settings remained constant. Respiratory rate of 30 breaths/min and tidal volume equal to FRC were chosen because in previous studies they proved sufficient to reach a sublethal mechanical power potentially associated with lung damage.2 

Experimental Procedures

The piglets were randomly allocated to six PEEP groups (0, 4, 7, 11, 14, and 18 cm H2O) of six animals each. FRC was measured with the helium-dilution technique8  during muscle relaxation and after lung recruitment at baseline, 24 h, and 48 h.

The animals were ventilated for a targeted 50 h at the randomly assigned PEEP level. Respiratory mechanics, gas exchange, and hemodynamic variables were assessed every 6 h. The animals were euthanized at the end of the experiment and autopsied. Lung tissue samples were collected for histologic analysis and wet-to-dry ratio6  (fig. S1, Supplemental Digital Content, https://links.lww.com/ALN/B790). We defined the wet-to-dry index as follows: , where H1 is the actual hours of experiment, and 50 is the hours planned for the experiment. This ratio accounts for the shorter time available for edema formation in the animals that died before intended. The adjustment assumes that edema accumulates linearly with time (see Supplemental Digital Content for details, https://links.lww.com/ALN/B790).

Mechanical Power

The mechanical power was derived by multiplying each component of the motion equation by the minute ventilation.1  Accordingly, in the respiratory system:

where 0.098 is the conversion factor from L*cm H2O to J, RR is the respiratory rate (breaths per minute), is the tidal volume (l), ERS is the elastance of the respiratory system (cm H2O/l), I:E is the ratio between inspiratory and expiratory time, and RRS is the airway and tissue resistances (cm H2O · l−1 · s−1). Assuming a constant rate of inflation, the term is the energy required to inflate the respiratory system (i.e., the product equals the driving pressure). The term is the energy required to overcome the airways and tissue resistances (i.e., the resistive power, when related to time). The term is the energy required to equilibrate the potential energy stored in the system at PEEP level (i.e., PEEP-related mechanical power, when related to time).

The mechanical power applied to the lung was computed as follows:

where EL is the lung elastance and the term is 1/inspiratory time. The lung tissue resistances, which cannot be assessed during inflation, can be estimated during lung stress relaxation after end-expiratory occlusion. Thus, the tissue resistance estimate is the quotient of (PL1 − PL2)/inspiratory flow, where PL2 is the transpulmonary pressure at the end of the plateau, and PL1 is the transpulmonary pressure at the first time point of flow cessation. The product ΔV2 × lung tissue resistances × 1/TI has the dimensions of pressure times volume and represents the energy spent to overcome the lung tissue resistances. PawPEEP and PawZEEP are the airway pressures measured at PEEP and zero end-expiratory pressure (ZEEP), respectively; PesPEEP and PesZEEP are the esophageal pressures measured at PEEP and ZEEP, respectively. Driving pressure-related, resistive, and PEEP-related lung power mirror the components described above for the whole respiratory system. The power time point used relative to outcome was the one measured at baseline.

Statistical Analysis

No a priori statistical power calculation was conducted, and the sample size was based on our experience from previous experiments. The data are presented as means ± SD unless otherwise specified. Student’s t test was used to compare lower versus higher PEEP in table 1. One-way ANOVA was used to evaluate differences among groups at baseline and at the end of the experiment and to compare different groups of PEEP. To evaluate the differences among groups over time, we used a linear mixed effects model with PEEP group, time in hours, and their interactions as fixed effects and animals as a random variable. This model ignores the missing data but generates outputs with the same numbers of observations compared with the original data set. Post hoc comparisons were conducted using the Tukey test. Correlation analysis was made with the Pearson method. We used Kaplan–Meier curves and log-rank tests to examine differences in morality across time. A P value < 0.05 was considered statistically significant (two-tailed testing). Outliers were not excluded from the analyses. To identify a possible “damaging” mechanical power threshold, we constructed receiver operating characteristic curves using the medians of the various variables as cutoffs. A logistic regression was used to identify which variables were independently associated with a lung mechanical power threshold. Analyses were performed with R software (R Project for Statistical Computing, https://www.r-project.org/).

Table 1.

Lung Weights and Wet-to-Dry Values

Lung Weights and Wet-to-Dry Values
Lung Weights and Wet-to-Dry Values

Positive End-expiratory Pressure and Mechanical Power

In figure 1, we show the mechanical power applied to the lungs and respiratory system throughout the experiment as a function of the set PEEP (fig. 1, A and B). Each stacked column shows the different components of the mechanical power. The total mechanical power applied to the lungs was similar at PEEP levels of 0, 4, and 7 cm H2O (P = 0.513), because the increase caused by rising PEEP was offset by the simultaneous decreases in driving pressure and the resistive components. In contrast, at PEEP 11, 14, and 18 cm H2O, the mechanical power increased proportionally to the applied PEEP, whereas the dynamic and resistive components of power remained unchanged. Of note, the driving pressure-related power significantly decreased from ZEEP to PEEP 4 and 7 cm H2O (lower PEEP groups), whereas it increased to levels similar to ZEEP in the higher PEEP groups. Therefore, the driving pressure-related power was significantly lower at PEEP 4 and 7 cm H2O and similar at ZEEP and at the higher PEEP levels of 11, 14, and 18 cm H2O.

Fig. 1.

Column chart showing the mechanical power applied to the respiratory system (top) and to the lung (bottom) at set positive end-expiratory pressure (PEEP) values of 0, 4, 7, 11, 14, and 18 cm H2O. Note that power is scaled differently for respiratory system (0 to 45 J/min) and the lung (0 to 25 J/min). ***P <0.001, driving pressure related power PEEP 0 vs. PEEP 4. §§P <0.05, driving pressure related lung power PEEP 0 versus PEEP 7.

Fig. 1.

Column chart showing the mechanical power applied to the respiratory system (top) and to the lung (bottom) at set positive end-expiratory pressure (PEEP) values of 0, 4, 7, 11, 14, and 18 cm H2O. Note that power is scaled differently for respiratory system (0 to 45 J/min) and the lung (0 to 25 J/min). ***P <0.001, driving pressure related power PEEP 0 vs. PEEP 4. §§P <0.05, driving pressure related lung power PEEP 0 versus PEEP 7.

Close modal

Mechanical power applied to the lung and to the respiratory system increased in each PEEP group over time, with a more pronounced rise after 24 to 30 h (fig. S2, A and B, Supplemental Digital Content, https://links.lww.com/ALN/B790). The different rates at which lung power increased in the higher and lower PEEP level groups are revealed by the different slopes of cumulative lung energy versus time (fig. S2C, Supplemental Digital Content, https://links.lww.com/ALN/B790). Of the 36 animals, 5 died during the experiment from either tension pneumothorax (n = 1; PEEP 18 cm H2O) or hemodynamic collapse (n = 2, PEEP 14 cm H2O; n = 2, PEEP 18 cm H2O). The difference in mortality among the PEEP groups was significant (P = 0.012; fig. S3, Supplemental Digital Content, https://links.lww.com/ALN/B790).

PEEP Effects on Lung Mechanics

At baseline (time 0), plateau airway pressure, driving airway pressure, and driving transpulmonary pressure were all correlated with the end-expiratory lung volume (FRC + PEEP volume; P < 0.001; r2 plateau airway pressure, 0.85; r2 driving airway pressure, 0.50; r2 driving transpulmonary pressure, 0.24; data not shown). The time course of these variables is displayed in fig. S4, Supplemental Digital Content, https://links.lww.com/ALN/B790). As shown, they all increased in every group, especially after 24 to 30 h. Of note, the greatest deteriorations of both airway and transpulmonary driving pressures were observed in the ZEEP group (see also table S1, Supplemental Digital Content, https://links.lww.com/ALN/B790).

These increases in pressure were uniquely due to the worsening of lung elastance, which accounted for an overall increase in the respiratory system elastance despite a small decrease in the chest wall elastance (fig. 2). All baseline elastance values were positively correlated with the baseline end-expiratory lung volume (P < 0.001; r2 lung elastance, 0.368; r2 respiratory system elastance, 0.523; r2 chest-wall elastance, 0.285; data not shown), and the greatest worsening of lung elastance was observed in the six animals treated at ZEEP (from 23 ± 4.8 to 46 ± 23 cm H2O/l; P = 0.044; fig. 2B). Lung stress at baseline was significantly higher in the groups with higher PEEP (table S1) and tended to increase during the experiment, particularly after 24 to 30 h (fig. S5, Supplemental Digital Content, https://links.lww.com/ALN/B790). In contrast, the associated strain did not change significantly throughout the experiment. Consequently, the specific elastance, which is the proportionality constant between stress and strain that reflects the intrinsic elasticity of lung parenchyma, worsened with time both in the lower and the higher PEEP groups. The time courses of these variables are presented in figure S5 in the Supplemental Digital Content (see also table S1, Supplemental Digital Content, https://links.lww.com/ALN/B790).

Fig. 2.

Time course of respiratory system elastance (linear mixed effects model: time, P < 0.001; positive end-expiratory pressure [PEEP], P < 0.001; PEEP:time interaction, P = 0.445; left), lung elastance (linear mixed effects model: time, P < 0.001; PEEP, P = 0.002; PEEP:time interaction, P = 0.499; middle), and chest wall elastance (linear mixed effects model: time, P = 0.005; PEEP, P = 0.103; PEEP:time interaction, P = 0.125; right) in the different PEEP groups.

Fig. 2.

Time course of respiratory system elastance (linear mixed effects model: time, P < 0.001; positive end-expiratory pressure [PEEP], P < 0.001; PEEP:time interaction, P = 0.445; left), lung elastance (linear mixed effects model: time, P < 0.001; PEEP, P = 0.002; PEEP:time interaction, P = 0.499; middle), and chest wall elastance (linear mixed effects model: time, P = 0.005; PEEP, P = 0.103; PEEP:time interaction, P = 0.125; right) in the different PEEP groups.

Close modal

PEEP Effects on Gas Exchange

The Pao2/Fio2 ratio was different among the PEEP groups and decreased slightly but significantly over the course of the experiment (P < 0.001), especially at ZEEP (from 632 ± 59 to 505 ± 106 mmHg; P = 0.028). The Pao2/Fio2 ratio paralleled the shunt fraction behaviors (fig. S6, Supplemental Digital Content, https://links.lww.com/ALN/B790, for the time course). The ventilatory protocol led to hypocapnia in all animals, both at baseline and during the course of the experiment (fig. S6C, Supplemental Digital Content, https://links.lww.com/ALN/B790). Dead space increased in all groups, especially in the PEEP 18 cm H2O group (P < 0.001; fig. S6D, Supplemental Digital Content, https://links.lww.com/ALN/B790). Base excess and pH were similar at baseline in all groups and steadily decreased in every group during the course of the experiment, even though lactate levels remained more or less constant. The time course of arterial base excess and pH is shown in figure S7, Supplemental Digital Content (see also table S2, Supplemental Digital Content, https://links.lww.com/ALN/B790).

Hemodynamics

As shown in figure 3 and table S3 (Supplemental Digital Content, https://links.lww.com/ALN/B790), cardiac output (fig. 3A) was similar at baseline in all groups and decreased significantly over time (P < 0.001). Mean arterial pressure (fig. 3B) differed significantly between the PEEP groups at baseline (P = 0.045). It decreased significantly over time, but there was no identifiable pattern associated with the PEEP levels. On the other hand, the volume of infused fluids required to maintain perfusion was proportional to the level of PEEP at baseline and throughout the experiment (fig. 3C). The noradrenaline infusion rate required to maintain arterial pressure was proportional to the PEEP level, but little was needed in the ZEEP and the 4 cm H2O PEEP groups (fig. 3D).

Fig. 3.

Time course of cardiac output (linear mixed effects model: time, P < 0.001; positive end-expiratory pressure [PEEP], P = 0.684; PEEP:time interaction, P = 0.315; A), mean arterial pressure (linear mixed effects model: time, P < 0.001; PEEP, P = 0.543; PEEP:time interaction, P = 0.011; B), cumulative resuscitation fluids (linear mixed effects model: time, P < 0.001; PEEP, P < 0.001; PEEP:time interaction, P < 0.001; C), and noradrenaline (linear mixed effects model: time, P < 0.001; PEEP P = 0.001; PEEP:time interaction, P = 0.003; D) in the different PEEP groups.

Fig. 3.

Time course of cardiac output (linear mixed effects model: time, P < 0.001; positive end-expiratory pressure [PEEP], P = 0.684; PEEP:time interaction, P = 0.315; A), mean arterial pressure (linear mixed effects model: time, P < 0.001; PEEP, P = 0.543; PEEP:time interaction, P = 0.011; B), cumulative resuscitation fluids (linear mixed effects model: time, P < 0.001; PEEP, P < 0.001; PEEP:time interaction, P < 0.001; C), and noradrenaline (linear mixed effects model: time, P < 0.001; PEEP P = 0.001; PEEP:time interaction, P = 0.003; D) in the different PEEP groups.

Close modal

Mean pulmonary pressure and wedge pressures increased both with increasing PEEP and over time (fig. S8, Supplemental Digital Content, https://links.lww.com/ALN/B790). Similar changes were seen in the fluid balance (fig. S9, Supplemental Digital Content, https://links.lww.com/ALN/B790). Arterial lactate levels slightly increased over time but were not different among the treatment groups (see also fig. S10 and table S3, Supplemental Digital Content, https://links.lww.com/ALN/B790).

PEEP and Edema

Table 1 (see also fig. S11, Supplemental Digital Content, https://links.lww.com/ALN/B790) displays the lung weight at the end of the experiment as well as the wet-to-dry indexes of lung, liver, bowel, kidney, and skeletal muscle. The average lung weight in these 36 piglets was significantly higher than the normal lung weight we previously measured by computed tomography scan in 73 healthy piglets of the same body weight (371 ± 105 g vs. 321 ± 40 g; P < 0.001),5  and it did not differ significantly between the groups. However, the wet-to-dry index, which accounts for the experimental duration, slightly increased with rising PEEP, reaching statistical significance (P = 0.006). The wet-to-dry index of all studied organs was higher among animals ventilated with higher PEEP, although the difference was only statistically significant for the lungs. This may have been due to an inadequate statistical power because of the small number of samples available for measuring the wet-to-dry index of liver, kidney, bowel, and muscle (11 and 10 in the higher and lower PEEP groups, respectively).

Lung Histology

We found no differences, either between left and right lungs or among different lung regions, i.e., apex, middle, and basal. The lesions in these prone animals were also uniformly distributed between dependent and nondependent regions. We therefore analyzed these histologic findings together, regardless of their anatomical position. A representative overview of the macroscopic aspect of the lungs in different PEEP groups is available in supplemental figure S12 (Supplemental Digital Content, https://links.lww.com/ALN/B790). Histologic data are reported in figure 4 and supplemental figure S13 (Supplemental Digital Content, https://links.lww.com/ALN/B790). As shown in the upper panel of figure 4, the most frequent finding was vascular congestion, followed by inflammatory cell infiltration, alveolar collapse/atelectasis, and septal dilatation/rupture. A similar pattern recurred with each of these lesions. Indeed, most of them were highly represented at ZEEP. They all decreased in the animals treated at 4 cm H2O but increased at higher PEEP levels. The decrease from ZEEP to PEEP 4 was significant for vascular congestion (P < 0.001), inflammation (P < 0.001), atelectasis (P < 0.001), septal rupture (P = 0.041), emphysema-like lesions (P < 0.001), and intravascular thrombi (P < 0.001). The incidence of vascular congestion, inflammation, atelectasis, and septal rupture increased significantly from PEEP 4 to PEEP 14 to 18 cm H2O (P < 0.001). The behavior of the other, less frequent lesions is reported in the supplement (fig. S13, Supplemental Digital Content, https://links.lww.com/ALN/B790). Briefly, at higher PEEP (14 to 18 cm H2O), alveolar edema was lower, and alveolar hemorrhage was higher than at lower PEEP. No hyaline membranes were observed.

Fig. 4.

(Upper) Column chart showing the percentage of optical field affected by the different histopathologic lesions in all the experimental animals. (Lower) Percentages of optical field with vascular congestion (A), inflammatory infiltrates (B), alveolar collapse/atelectasis (C), and septal dilations/ruptures (D) as a function of experimental positive end-expiratory pressure (PEEP) groups. Significant statistical differences are indicated by black circles (zero end-expiratory pressure), white circles (4 cm H2O), section signs (§) (7 cm H2O), and white triangles (11 cm H2O).

Fig. 4.

(Upper) Column chart showing the percentage of optical field affected by the different histopathologic lesions in all the experimental animals. (Lower) Percentages of optical field with vascular congestion (A), inflammatory infiltrates (B), alveolar collapse/atelectasis (C), and septal dilations/ruptures (D) as a function of experimental positive end-expiratory pressure (PEEP) groups. Significant statistical differences are indicated by black circles (zero end-expiratory pressure), white circles (4 cm H2O), section signs (§) (7 cm H2O), and white triangles (11 cm H2O).

Close modal

PEEP and Mechanical Power Thresholds

The variables suggestive of lung damage are presented in table 2. The measures obtained in all piglets—regardless of allocation group—were pooled, and their median values were computed. We then computed—as “threshold”—the mechanical power associated with the median value of each variable. Therefore, this threshold represents the mechanical power that is associated with a damage greater or lower than the 50% median value. As shown, the mechanical power thresholds averaged were 13 ± 1 J/min for the lung and 25 ± 1.7 J/min for the respiratory system. We introduced into a logistic regression model the lung-specific elastance, dead space, wet-to-dry index, noradrenaline requirement, and pulmonary artery mean pressure. These variables were selected as representative of lung mechanics, gas exchange, anatomy, and hemodynamics. The specific elastance odds ratio was 1.83 (CI, 1.6 to 2.2), the pulmonary artery pressure odds ratio was 1.13 (CI, 1.1 to 1.2), and the noradrenaline odds ratio was 565 (CI, 66 to 7,749). These three variables were significantly related to the lung mechanical power threshold of 13 J/min (P < 0.001 for all). The dead space and the wet-to-dry index were not significantly related with the mechanical power threshold.

Table 2.

Summary of Findings of the Receiver Operating Characteristic Curves

Summary of Findings of the Receiver Operating Characteristic Curves
Summary of Findings of the Receiver Operating Characteristic Curves

Summary of Results

The mean values (± SDs) of each of the variables considered above are reported for each group in the supplement: lung and respiratory system mechanics variables (Table S1, Supplemental Digital Content, https://links.lww.com/ALN/B790), gas exchange variables (table S2, Supplemental Digital Content, https://links.lww.com/ALN/B790), hemodynamic variables (table S3, Supplemental Digital Content, https://links.lww.com/ALN/B790), and intergroup comparisons (table S4, Supplemental Digital Content, https://links.lww.com/ALN/B790).

If we strictly consider the effects of the presence of positive pressure at end expiration (isolating PEEP itself as a potentially damaging contributor), this study shows that its increase from 4 to 7, 11, 14, and 18 cm H2O is associated with proportionally greater damage to the previously healthy lung, whether considering gross anatomy, histology, hemodynamics, or the overall clinical scenario. Indeed, the total energy per cycle required to overcome the presence of PEEP contributed to the lung damage. On the other hand, under unchanging ventilating conditions, it is evident that, over time, the complete absence of PEEP harmed the initially healthy lungs of these anesthetized animals. Although such damages were almost fully prevented at 4 cm H2O PEEP and partially prevented at 7 cm H2O, they were similar and even higher than at ZEEP at PEEP values of 11, 14, or 18 cm H2O (fig. 4).

In keeping constant the tidal volume and the respiratory rate, the total mechanical power theoretically should have increased linearly with PEEP.1  However, we found it similar at 0, 4, and 7 cm H2O because of the improving respiratory system elastance over that range. Indeed, up to 7 cm H2O, the increases in the PEEP component of delivered mechanical power were offset by parallel decreases of the driving pressure component. We may envisage several hypotheses for the presence of damage of different extents at the same total mechanical power level. First, it is possible that, for the same “package” of mechanical power, the contribution of its components does not carry the same weight in inducing damage. A possible explanation could be that the higher driving pressure at ZEEP could have played a greater role than the PEEP increase. Alternatively, staying more strictly within the mechanical power hypothesis, it is possible that the decrease in FRC at ZEEP caused by anesthesia and paralysis (prevented by PEEP) resulted in an increased specific power, i.e., the power normalized for the dimension of the lung to which it is applied. In addition, it is possible that the power focused at the interface between units undergoing collapse and decollapse is far greater than in the remaining lung parenchyma.

Indeed, in the ZEEP animals, lung damage manifested as decreased FRC, increased atelectasis, and inflammatory infiltrates, as well as increased lung elastance and specific elastance. Of note, the ZEEP animals had the lowest lung weight and wet-to-dry index, noradrenaline infusion rates, and resuscitation fluid volumes, whereas the mean pulmonary artery and wedge pressures were similar to the ones measured in piglets treated at PEEP 4 or 7 cm H2O. The bulk of data suggests that, at ZEEP, atelectrauma with increased driving pressure/increased focused mechanical power is the most relevant cause of lung damage in our model, occurring without major hemodynamic impairment.

By increasing PEEP from 4 to 11, 14, or 18 cm H2O, we observed progressively greater damage, and we measured a substantial increase in lung mechanical power. This increase was mainly related to the increase in the PEEP-related power component, whereas the component related selectively to driving pressure (excluding PEEP) was similar to that applied at ZEEP. Some of the damages observed at ZEEP, such as increased lung weight, histologic atelectasis, and inflammatory infiltrates, were not prevented by these higher levels of PEEP, whereas other indicators, namely stress, strain, lung elastance, and specific lung elastance, even increased significantly by 40 to 70%. Furthermore, new alterations manifested at these higher PEEP levels, such as a striking increase in mean pulmonary artery pressure (by 40%), additional fluid requirements (by 145%), and greater needs for noradrenaline infusion (by 100%). Finally, the death of five animals treated at 14 to 18 cm H2O PEEP was clearly attributable to hemodynamic impairment; one of these deaths was attributable to tension pneumothorax. Indeed, increasing PEEP above 7 cm H2O did not prevent lung damage and added a relevant burden of hemodynamic instability that led to significantly increased damage and mortality.

In these experiments, we identified a threshold level of lung mechanical power of 13 J/min that split anatomical, mechanical, gas exchange, and hemodynamic damage variables at their median values. This is similar to the threshold value of 12.5 J/min that we found in a different experimental setting, where the mechanical power was modified by changing the respiratory frequency while keeping the tidal volume constant at ZEEP.3  It must be pointed out, however, that such a threshold should be considered only as the limit above which mechanical ventilation may be lethal and not as the limit below which mechanical ventilation will safely maintain the lungs normally. Actually, we observed lung damages in all PEEP groups, likely because the applied ventilation was deleterious, regardless of the PEEP level.

One of the main conceptual challenges of this experiment was how to assess ventilator-induced lung injury. Ventilator-induced lung injury encompasses a variety of anatomical, physiologic, and clinical signs.9,10  Experimentally, the most common findings of atelectrauma are increased lung weight (edema), shunt, dead space, and elastance, as well as histologic atelectasis and inflammatory infiltrates.10  The most typical feature of volutrauma, however, is hyperinflation, with its effects on lung mechanics and hemodynamics. In our study, atelectrauma was the most likely cause of damage manifested at ZEEP, even though edema was not an important feature, probably because mean airway pressure was sufficiently high to prevent it.11–13  Increasing PEEP moved the pattern toward volutrauma, with its typical effects on lung mechanics, hemodynamics, and mortality, but without gas exchange impairment or relevant edema, as indicated by less alveolar and perivascular edema observed at the highest PEEP levels (fig. S13, B and C, fig. S13, Supplemental Digital Content, https://links.lww.com/ALN/B790). Of note, the putative increase or decrease of ventilator-induced lung injury in clinical trials is measured as increased or decreased mortality rate,14–18  although the mortality attributable to mechanical ventilation or ventilator-induced lung injury per se is far from defined. In contrast, in our model, all negative effects, including mortality, could be solely attributed to the consequences of mechanical ventilation, because the experiment started with healthy lungs of animals supported in their naturally prone position.

In this study, we wanted to investigate the role of PEEP as a component of mechanical power. Unfortunately, it is not possible (without artificial lungs) to increase the mechanical power actually applied selectively by PEEP, i.e., without accompanying changes in some of the other contributing components. In our model that mandated constant tidal volume and frequency, an increase in PEEP unavoidably may result in linear increases in plateau and driving pressures while tracking along the straight part of the volume–pressure curve, but in their exponential increases while tracking along the uppermost portion of that curve. The issue, however, is not whether PEEP or tidal volume or driving pressure or frequency is a greater or lesser inducer of ventilator-induced lung injury but rather how to consider them together, because ventilator-induced lung injury depends on how all these variables are set when ventilating a given patient. We believe that, at this stage, the mechanical power is a unifying tool that reminds us that all its components should be considered when attempting to prevent ventilator-induced lung injury. Indeed, PEEP may be an important determinant of damage in some settings (as explored in this study), whereas in others it would be expected to be a minor contributor compared with the dominating influence of the driving pressure,19  respiratory frequency,3  or their combination20  (depending on their magnitudes and interactions).

Limitations

Our experiments have several limitations, because they did not answer some fundamental questions related to power. First, as discussed above, the increase in PEEP at the same tidal volume is associated with increase in plateau pressure and driving pressure, in a sense making arbitrary the interpretation of the results. Second, we do not know how to best normalize the mechanical power. This is not really relevant in our experiment, because the inherent normalization was provided by the use of healthy animals of similar weight and age. The normalization would be of paramount importance when comparing different species. The influence of species on damaging power threshold for the lung is also unknown, and the threshold value of 13 J/min that we found here and previously in piglets cannot be translated with quantitative certainty to the human being. We should consider, in addition, that stresses related to PEEP levels of 4 and 7 cm H2O in piglets are roughly equivalent to 8 and 14 cm H2O in humans, because of the differences in specific lung elastance (~6 vs. ~12 cm H2O).5  Furthermore, we do not clearly know how mechanical power is distributed in space, (i.e., “baby lung” size and stress-raising inhomogeneity21 ), and in time, i.e., within the phases of the respiratory cycle (inspiration vs. expiration). We also do not know whether all “packages” of mechanical power result in equivalent damage or whether power-related damage depends critically upon the composition of its determinants, as suggested by the comparison of ZEEP and 4 to 7 cm H2O PEEP. Obviously, if a mechanical power threshold for damage were to be identified in humans, it would represent a boundary limit beyond which the applied pattern of mechanical ventilation is not acceptable and therefore represent a key indicator of the need for alternative methods of ventilatory support.

This work was made possible by the generous donation of Charlotte Munz, Göttingen, Germany, to the Department of Anesthesiology, University of Göttingen, Göttingen, Germany.

Support was provided solely from institutional and/or departmental sources.

The authors declare no competing interests.

1.
Gattinoni
L
,
Tonetti
T
,
Cressoni
M
,
Cadringher
P
,
Herrmann
P
,
Moerer
O
,
Protti
A
,
Gotti
M
,
Chiurazzi
C
,
Carlesso
E
,
Chiumello
D
,
Quintel
M
:
Ventilator-related causes of lung injury: The mechanical power.
Intensive Care Med
2016
;
42
:
1567
75
2.
Protti
A
,
Cressoni
M
,
Santini
A
,
Langer
T
,
Mietto
C
,
Febres
D
,
Chierichetti
M
,
Coppola
S
,
Conte
G
,
Gatti
S
,
Leopardi
O
,
Masson
S
,
Lombardi
L
,
Lazzerini
M
,
Rampoldi
E
,
Cadringher
P
,
Gattinoni
L
:
Lung stress and strain during mechanical ventilation: Any safe threshold?
Am J Respir Crit Care Med
2011
;
183
:
1354
62
3.
Cressoni
M
,
Gotti
M
,
Chiurazzi
C
,
Massari
D
,
Algieri
I
,
Amini
M
,
Cammaroto
A
,
Brioni
M
,
Montaruli
C
,
Nikolla
K
,
Guanziroli
M
,
Dondossola
D
,
Gatti
S
,
Valerio
V
,
Vergani
GL
,
Pugni
P
,
Cadringher
P
,
Gagliano
N
,
Gattinoni
L
:
Mechanical power and development of ventilator-induced lung injury.
Anesthesiology
2016
;
124
:
1100
8
4.
Protti
A
,
Maraffi
T
,
Milesi
M
,
Votta
E
,
Santini
A
,
Pugni
P
,
Andreis
DT
,
Nicosia
F
,
Zannin
E
,
Gatti
S
,
Vaira
V
,
Ferrero
S
,
Gattinoni
L
:
Role of strain rate in the pathogenesis of ventilator-induced lung edema.
Crit Care Med
2016
;
44
:
e838
45
5.
Protti
A
,
Andreis
DT
,
Milesi
M
,
Iapichino
GE
,
Monti
M
,
Comini
B
,
Pugni
P
,
Melis
V
,
Santini
A
,
Dondossola
D
,
Gatti
S
,
Lombardi
L
,
Votta
E
,
Carlesso
E
,
Gattinoni
L
:
Lung anatomy, energy load, and ventilator-induced lung injury.
Intensive Care Med Exp
2015
;
3
:
34
6.
Protti
A
,
Andreis
DT
,
Monti
M
,
Santini
A
,
Sparacino
CC
,
Langer
T
,
Votta
E
,
Gatti
S
,
Lombardi
L
,
Leopardi
O
,
Masson
S
,
Cressoni
M
,
Gattinoni
L
:
Lung stress and strain during mechanical ventilation: Any difference between statics and dynamics?
Crit Care Med
2013
;
41
:
1046
55
7.
Webb
HH
,
Tierney
DF
:
Experimental pulmonary edema due to intermittent positive pressure ventilation with high inflation pressures: Protection by positive end-expiratory pressure.
Am Rev Respir Dis
1974
;
110
:
556
65
8.
Chiumello
D
,
Cressoni
M
,
Chierichetti
M
,
Tallarini
F
,
Botticelli
M
,
Berto
V
,
Mietto
C
,
Gattinoni
L
:
Nitrogen washout/washin, helium dilution and computed tomography in the assessment of end expiratory lung volume.
Crit Care
2008
;
12
:
1
8
9.
Kumar
A
,
Pontoppidan
H
,
Falke
KJ
,
Wilson
RS
,
Laver
MB
:
Pulmonary barotrauma during mechanical ventilation.
Crit Care Med
1973
;
1
:
181
6
10.
Dreyfuss
D
,
Soler
P
,
Basset
G
,
Saumon
G
:
High inflation pressure pulmonary edema: Respective effects of high airway pressure, high tidal volume, and positive end-expiratory pressure.
Am Rev Respir Dis
1988
;
137
:
1159
64
11.
Colmenero-Ruiz
M
,
Fernández-Mondéjar
E
,
Fernández-Sacristán
MA
,
Rivera-Fernández
R
,
Vazquez-Mata
G
:
PEEP and low tidal volume ventilation reduce lung water in porcine pulmonary edema.
Am J Respir Crit Care Med
1997
;
155
:
964
70
12.
Myers
JC
,
Reilley
TE
,
Cloutier
CT
:
Effect of positive end-expiratory pressure on extravascular lung water in porcine acute respiratory failure.
Crit Care Med
1988
;
16
:
52
4
13.
Russell
JA
,
Hoeffel
J
,
Murray
JF
:
Effect of different levels of positive end-expiratory pressure on lung water content.
J Appl Physiol Respir Environ Exerc Physiol
1982
;
53
:
9
15
14.
Cavalcanti
AB
,
Suzumura
ÉA
,
Laranjeira
LN
,
Paisani
DM
,
Damiani
LP
,
Guimarães
HP
,
Romano
ER
,
Regenga
MM
,
Taniguchi
LNT
,
Teixeira
C
,
Pinheiro de Oliveira
R
,
Machado
FR
,
Diaz-Quijano
FA
,
Filho
MSA
,
Maia
IS
,
Caser
EB
,
Filho
WO
,
Borges
MC
,
Martins
PA
,
Matsui
M
,
Ospina-Tascón
GA
,
Giancursi
TS
,
Giraldo-Ramirez
ND
,
Vieira
SRR
,
Assef
MDGPL
,
Hasan
MS
,
Szczeklik
W
,
Rios
F
,
Amato
MBP
,
Berwanger
O
,
Ribeiro de Carvalho
CR
;
Writing Group for the Alveolar Recruitment for Acute Respiratory Distress Syndrome Trial (ART) Investigators
:
Effect of lung recruitment and titrated positive end-expiratory pressure (PEEP) vs. low PEEP on mortality in patients with acute respiratory distress syndrome: A randomized clinical trial.
JAMA
2017
;
318
:
1335
45
15.
Acute Respiratory Distress Syndrome Network
:
Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome.
N Engl J Med
2000
;
342
:
1301
8
16.
Meade
MO
,
Cook
DJ
,
Guyatt
GH
,
Slutsky
AS
,
Arabi
YM
,
Cooper
DJ
,
Davies
AR
,
Hand
LE
,
Zhou
Q
,
Thabane
L
,
Austin
P
,
Lapinsky
S
,
Baxter
A
,
Russell
J
,
Skrobik
Y
,
Ronco
JJ
,
Stewart
TE
;
Lung Open Ventilation Study Investigators
:
Ventilation strategy using low tidal volumes, recruitment maneuvers, and high positive end-expiratory pressure for acute lung injury and acute respiratory distress syndrome: A randomized controlled trial.
JAMA
2008
;
299
:
637
45
17.
Brower
RG
,
Lanken
PN
,
MacIntyre
N
,
Matthay
MA
,
Morris
A
,
Ancukiewicz
M
,
Schoenfeld
D
,
Thompson
BT
;
National Heart, Lung, and Blood Institute ARDS Clinical Trials Network
:
Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome.
N Engl J Med
2004
;
351
:
327
36
18.
Mercat
A
,
Richard
JC
,
Vielle
B
,
Jaber
S
,
Osman
D
,
Diehl
JL
,
Lefrant
JY
,
Prat
G
,
Richecoeur
J
,
Nieszkowska
A
,
Gervais
C
,
Baudot
J
,
Bouadma
L
,
Brochard
L
;
Expiratory Pressure (Express) Study Group
:
Positive end-expiratory pressure setting in adults with acute lung injury and acute respiratory distress syndrome: A randomized controlled trial.
JAMA
2008
;
299
:
646
55
19.
Amato
MB
,
Barbas
CS
,
Medeiros
DM
,
Magaldi
RB
,
Schettino
GP
,
Lorenzi-Filho
G
,
Kairalla
RA
,
Deheinzelin
D
,
Munoz
C
,
Oliveira
R
,
Takagaki
TY
,
Carvalho
CR
:
Effect of a protective-ventilation strategy on mortality in the acute respiratory distress syndrome.
N Engl J Med
1998
;
338
:
347
54
20.
Ferguson
ND
,
Cook
DJ
,
Guyatt
GH
,
Mehta
S
,
Hand
L
,
Austin
P
,
Zhou
Q
,
Matte
A
,
Walter
SD
,
Lamontagne
F
,
Granton
JT
,
Arabi
YM
,
Arroliga
AC
,
Stewart
TE
,
Slutsky
AS
,
Meade
MO
;
OSCILLATE Trial Investigators; Canadian Critical Care Trials Group
:
High-frequency oscillation in early acute respiratory distress syndrome.
N Engl J Med
2013
;
368
:
795
805
21.
Cressoni
M
,
Cadringher
P
,
Chiurazzi
C
,
Amini
M
,
Gallazzi
E
,
Marino
A
,
Brioni
M
,
Carlesso
E
,
Chiumello
D
,
Quintel
M
,
Bugedo
G
,
Gattinoni
L
:
Lung inhomogeneity in patients with acute respiratory distress syndrome.
Am J Respir Crit Care Med
2014
;
189
:
149
58