Your privacy, your choice

We use essential cookies to make sure the site can function. We also use optional cookies for advertising, personalisation of content, usage analysis, and social media.

By accepting optional cookies, you consent to the processing of your personal data - including transfers to third parties. Some third parties are outside of the European Economic Area, with varying standards of data protection.

See our privacy policy for more information on the use of your personal data.

for further information and to change your choices.
Skip to main content
Download PDF

Individualized PEEP can improve both pulmonary hemodynamics and lung function in acute lung injury

Critical Care volume 29, Article number: 107 (2025) Cite this article

Abstract

Rationale

There are several approaches to select the optimal positive end-expiratory pressure (PEEP), resulting in different PEEP levels. The impact of different PEEP settings may extend beyond respiratory mechanics, affecting pulmonary hemodynamics.

Objectives

To compare PEEP levels obtained with three titration strategies—(i) highest respiratory system compliance (CRS), (ii) electrical impedance tomography (EIT) crossing point; (iii) positive end-expiratory transpulmonary pressure (PL)—in terms of regional respiratory mechanics and pulmonary hemodynamics.

Methods

Experimental studies in two porcine models of acute lung injury: (I) bilateral injury induced in both lungs, generating a highly recruitable model (n = 37); (II) asymmetrical injury, generating a poorly recruitable model (n = 13). In all experiments, a decremental PEEP titration was performed monitoring PL, EIT (collapse, overdistention, and regional ventilation), respiratory mechanics, and pulmonary and systemic hemodynamics.

Measurements and main results

PEEP titration methods resulted in different levels of median optimal PEEP in bilateral lung injury: 14(12–14) cmH2O for CRS, 11(10–12) cmH2O for EIT, and 8(8–10) cmH2O for PL, p < 0.001. Differences were less pronounced in asymmetrical lung injury. PEEP had a quadratic U-shape relationship with pulmonary artery pressure (R2 = 0.94, p < 0.001), right-ventricular systolic transmural pressure, and pulmonary vascular resistance. Minimum values of pulmonary vascular resistance were found around individualized PEEP, when ventilation distribution and pulmonary circulation were simultaneously optimized.

Conclusions

In porcine models of acute lung injury with variable lung recruitability, both low and high levels of PEEP can impair pulmonary hemodynamics. Optimized ventilation and hemodynamics can be obtained simultaneously at PEEP levels individualized based on respiratory mechanics, especially by EIT and esophageal pressure.

Introduction

Setting positive end-expiratory pressure (PEEP) is a crucial aspect of mechanical ventilation and remains a challenge in clinical practice [1,2,3,4,5]. The optimal strategy for PEEP titration remains a topic of debate [6,7,8], and there is no universally accepted method. The PEEP-FiO2 table strategy [9] has been widely used because of its easy applicability, but as it is solely based on oxygenation, it has many drawbacks and poorly accounts for individual patient variability. In recent years, personalized approaches have gained attention, including the use of esophageal pressure monitoring [10], aiming for a slightly positive end-expiratory transpulmonary pressure (PL), and electrical impedance tomography (EIT) [11], targeting a compromise between lung overdistention and collapse (crossing point) [12]. The highest respiratory system compliance (CRS) is a classical individualized approach [13, 14], with the advantage that it could be determined without additional equipment other than the ventilator. It has been recently suggested that these approaches may result in very different levels of PEEP [15].

One primary goal for PEEP was to prevent alveolar collapse at the end of expiration and reverse severe hypoxemia [16]; however, it quickly became clear that PEEP could also result in alveolar overdistention and compromise hemodynamics, the latter resulting in a paradoxical benefit on oxygenation [17, 18]. Numerous experimental studies [5, 19,20,21,22], assessing the physiological effects of PEEP, have observed that both low and high levels of PEEP may be associated with ventilator-induced lung injury (VILI), and clinical trials have failed to prove a benefit from one approach versus the other on clinical outcomes [9, 23, 24]. A recent network metanalysis [25] has shown that the use of a higher PEEP (without prolonged lung recruitment strategy) was associated with lower risk of mortality than a lower PEEP strategy in patients with moderate to severe lung injury.

An often-overlooked aspect of PEEP titration is its impact on pulmonary hemodynamics, especially right ventricular afterload, and the interdependence between overall lung volume, blood gases and the overall resistance of pulmonary circulation. There is a classical relationship in physiology between pulmonary vascular resistance (PVR) and lung volumes, exhibiting a U-shape relationship, with the lowest PVR at the functional residual capacity (FRC) [26,27,28]. Based on this understanding, it has been assumed that the application of PEEP generally leads to a departure from FRC, ultimately resulting in an increase in PVR, as supported by earlier studies [29, 30]. A recent study [31] in 23 acute respiratory distress (ARDS) patients observed a negative impact of PEEP on PVR, especially in patients with low lung recruitability, indicated by recruitment-to-inflation ratio (R/I) < 0.5 (where PEEP is expected to create lung overdistention). In contrast, our previous experimental randomized trial suggested that pulmonary hemodynamics was commonly worse at low PEEP levels, especially when associated with massive lung collapse [5]. These contradictory findings raise concerns about our current PEEP titration strategies, highlighting the need to better understand the relationship between PEEP and pulmonary hemodynamics, especially during different degrees of lung injury—rather than in normal physiological conditions.

The main objectives of this study were 1) to estimate the PEEP level resulting from three different titration strategies, respectively using maximum CRS, EIT (crossing point), or PL (slightly positive) in two porcine models of acute lung injury with different recruitability (bilateral vs. asymmetrical); and 2) to assess the impact of PEEP on pulmonary hemodynamics in acutely injured lungs and compare the impact of the different titration methods on pulmonary circulation.

Methods

This is a secondary analysis of two series of experimental studies in two porcine models of acute lung injury [5, 32,33,34]. The study protocol was approved by the Animal Care Committee of the Peter Gilgan Centre for Research and Learning (PGRLC) at The Hospital for Sick Children – SickKids, reference number 1000058058. In this study, we included two series of experiments: I) Bilateral Lung Injury and II) Asymmetrical Lung Injury, as described in the following subsections.

Female Yorkshire pigs (35–50 kg) were sedated and paralyzed during the entire protocol. After intubation, pulmonary artery and femoral artery catheters (Swan-Ganz, Edwards Lifesciences, Irvine, United States, and PiCCO2, Getinge, Solna, Sweden), esophageal and gastric balloons (NutriVent®, Sidam, Italy), and EIT belt (PulmoVista® 500, Drager, Germany) were placed.

Experiment series I: bilateral lung injury

In the first series of experiments, after animal preparation, we induced lung injury in both lungs by a two-hit model (surfactant lavage and high stretch ventilation) [5].

After inducing lung injury, we performed a decremental PEEP titration from PEEP of 24–22 cmH2O to 4–0 cmH2O, by steps of 2 cmH2O (≥ 30 s/step), either on volume-controlled ventilation (with tidal volume of 6 mL/kg), when the EIT device was connected to the ventilator, or on pressure-controlled ventilation (with driving pressure of 15 cmH2O). During the decremental PEEP titration, we continuously recorded airway pressure and flow, esophageal pressure, systemic blood pressure, central venous pressure (CVP), and pulmonary artery pressure (PAP), using PowerLab and LabChart (ADInstruments, Dunedin, New Zealand). At each step of PEEP, we calculated CRS—as tidal volume divided by (plateau pressure minus total PEEP), end-expiratory PL—as airway pressure minus esophageal pressure (both at end expiration), mean arterial pressure (MAP), mean PAP, and right-ventricular (RV) systolic transmural pressure—as systolic PAP minus expiratory esophageal pressure. EIT data were also recorded continuously, and we measured the percentage of lung collapse, lung overdistention, and regional ventilation at each level of PEEP. Airway opening pressure and R/I ratio were estimated [35].

Optimal PEEP was titrated using three different methods:

  1. 1)

    EIT: optimal PEEP was identified at the crossing point between collapse and overdistention.

  2. 2)

    End-expiratory PL: optimal PEEP was defined as the lowest PEEP level at which end-expiratory PL was higher than zero.

  3. 3)

    CRS: optimal PEEP corresponded to the level that resulted in the highest CRS.

In two experiments of this series, we performed additional measurements to better understand the impact of PEEP and pulmonary hemodynamics on gas exchange and heart function. At each PEEP level, we recorded end-tidal carbon dioxide (EtCO2), collected arterial and mixed venous blood samples, and measured cardiac output (CO) and pulmonary capillary wedge pressure (PCWP). PVR was calculated as mean PAP minus PCWP divided by CO and multiplied by 80 [36].

Experiment series II: asymmetrical lung injury

In the second series of experiments, we induced lung injury in only one lung [33]. After animal preparation, selective lung intubation of the left lung was performed with a double lumen endotracheal tube with confirmation via EIT. Surfactant lavage and high stretch ventilation were performed only on the left lung, while the right lung was collapsed. After inducing lung injury, bilateral ventilation was restored with a single lumen endotracheal tube. Decremental PEEP titration and variables measurements were performed as in the first series of experiments. In nine experiments of this series, we also performed measurements of CO, PCWP, and PVR.

Data analysis and statistics

Categorical variables were expressed as percentage, and continuous variables were presented as mean and standard deviation or median and interquartile range according to their distribution. Statistical tests were two-tailed with α = 0.05, conducted using R (https://www.R-project.org/].

To compare the three levels of PEEP, we used repeated measures ANOVA and Duncan’s test. Correlations were assessed by Pearson’s test. Bland-Altman plots were built to estimate the mean difference between PEEP levels. Regression models were used to evaluate the relationship between PEEP and hemodynamic variables. To account for potential non-linear effects, we also explored quadratic regression models and evaluated the model fit using the Akaike information criterion (AIC). The dependent variable was pulmonary hemodynamics (i.e. mean PAP, RVTMP, or PVR), and the fixed effects included PEEP and PEEP squared. The intercept and the β coefficients for PEEP and PEEP squared were estimated, with standard errors (SE) provided for statistical inference. Linear mixed-effects model was employed to account for inter-case variability when analyzing the relationship between PEEP and mean PAP, and we calculated both the marginal and conditional R2 values. The marginal R2 represents the proportion of variance in the outcome variable explained by the fixed effects alone, while the conditional R2 reflects the variance explained by both the fixed effects and the random effects.

See extended methods in the supplements for more details.

Results

Experiment series I: bilateral lung injury

1) PEEP levels and regional compliance

Thirty-seven pigs with bilateral lung injury were included in the first series of experiments. After lung injury, mean PaO2/FiO2 was 123 ± 74 mmHg and CRS was 13 ± 3 mL/cmH2O at PEEP of 10 cmH2O. This model of lung injury was highly recruitable, with a median R/I ratio of 1.24(0.90–1.40).

The three strategies of PEEP titration performed in this study resulted in three different levels of optimal PEEP in this model (Fig. 1): median PEEP was 14(12–14) cmH2O at the highest CRS, 11(10–12) cmH2O at the EIT crossing point, and 8(8–10) cmH2O at slightly positive expiratory PL, p < 0.001 (Fig. 2A).

Fig. 1
figure 1

Positive end-expiratory pressure (PEEP) titration in pigs with bilateral lung injury (n = 37) according to three strategies. Black solid line represents the average electrical impedance tomography (EIT) crossing point between collapse and overdistension. Black dashed line represents the average highest respiratory system compliance (CRS). Black dotted line represents the average end-expiratory transpulmonary pressure (PL) slightly positive. Error bars represent standard error

Full size image
Fig. 2
figure 2

A The optimal positive end-expiratory pressure (PEEP) according to different PEEP titration strategies in pigs with bilateral lung injury (n = 37): CRS, highest respiratory system compliance; EIT, electrical impedance tomography crossing point; and PL, end-expiratory transpulmonary pressure slightly positive. * represents post-hoc p < 0.05. B Regional compliance in bilateral lung injury (n = 37). Black solid line represents the average EIT crossing point between collapse and overdistension. Black dashed line represents the average highest CRS. Black dotted line represents the average end-expiratory PL slightly positive. Error bars represent standard error

Full size image

There was a moderate correlation between CRS-PEEP and EIT-PEEP (Pearson's r = 0.59, p < 0.001), with a mean difference of 2 cmH2O (Supplemental Fig. 1A), and a weak, non-significant, correlation between CRS-PEEP and PL-PEEP (Pearson's r = 0.16, p = 0.345), with a mean difference of 4 cmH2O (Supplemental Fig. 1B).

In terms of regional respiratory system compliance (Fig. 2B), PEEP set based on CRS was almost a perfect reflection of the compliance of the dependent lung. The mean difference between ventral and dorsal ventilation was higher at PEEP set based on end-expiratory PL (18 ± 31%), than for both PEEP based on EIT (3 ± 19%, p = 0.008 and on CRS (0 ± 18%, p = 0.002), Supplemental Fig. 2.

2) Impact on pulmonary hemodynamics

We observed a quadratic relationship (U shape) between mean PAP and PEEP (Fig. 3A). The quadratic model (AIC = 29.7) provided a better fit to the data compared to the linear model (AIC = 56.6), as indicated by the lower AIC value. Similarly, the relationship between RV systolic transmural pressure and PEEP was also quadratic (Fig. 3B). To further investigate these relationships while accounting for inter-case variability, we performed a linear mixed-effects model, including a quadratic term, which confirmed the quadratic relationship (p < 0.001). The conditional R2 was 0.905, indicating that 90.5% of the variance in mean PAP can be explained by the model, while the marginal R2 was 0.041, reflecting the variance explained by the fixed effects alone. The β coefficients for PEEP and PEEP squared were − 1.18 (SE = 0.070, p < 0.001) and 0.053 (SE = 0.003, p < 0.001), respectively, with a random effect variance of 56.15 for the intercept and a residual variance of 6.21. These results demonstrate that mean PAP decreases with increasing PEEP up to a certain point, after which it begins to increase again, supporting the hypothesis of a quadratic (non-linear) relationship. We observed a negative linear correlation of mean systemic arterial pressure (conditional R2:0.88, p < 0.001) and CO (conditional R2:0.96, p < 0.001) with PEEP, a positive linear correlation between CVP and PEEP (conditional R2 = 0.97, p < 0.001), and a quadratic relationship between transmural CVP and PEEP (conditional R2 = 0.81) Supplemental Fig. 3.

Fig. 3
figure 3

Pulmonary hemodynamics in bilateral lung injury (n = 37): A Mean pulmonary artery pressure (Mean PAP) at each level of positive end-expiratory pressure (PEEP). B Right ventricular (RV) systolic transmural pressure at each level of PEEP. Dots represent mean and error bars represent standard error. Red line represents quadratic regression, and grey area shaded represents 95% confidence interval. Black solid line represents the average electrical impedance tomography (EIT) crossing point between collapse and overdistension. Black dashed line represents the average highest respiratory system compliance (CRS). Black dotted line represents the average end-expiratory transpulmonary pressure (PL) slightly positive

Full size image

Experiment series II: asymmetrical lung injury

1) PEEP levels and regional compliance

Thirteen pigs with asymmetrical lung injury were included in the second series of experiments. After lung injury, mean PaO2/FiO2 was 172 ± 48 mmHg and total CRS was 22 ± 4 mL/cm H2O at PEEP of 10 cmH2O. This model of lung injury had low-to-moderate lung recruitability, with a median R/I ratio of 0.67(0.29–0.98). The PEEP at slightly positive expiratory PL (6 ± 4 cmH2O) was lower than both PEEP at the EIT crossing point (9 ± 2 cmH2O, p = 0.047) and PEEP at the highest CRS (9 ± 3 cmH2O, p = 0.023), despite a noteworthy variability of distribution (Figs. 4 and 5A). There was a strong correlation between CRS-PEEP and EIT-PEEP (Pearson's r = 0.78, p = 0.002), with a mean difference of 0 cmH2O, and a weak, non-significant, correlation between CRS-PEEP and PL-PEEP (Pearson's r = 0.32, p = 0.282), with a mean difference of 4 cmH2O (Supplemental Fig. 4).

Fig. 4
figure 4

Positive end-expiratory pressure (PEEP) titration in pigs with asymmetrical lung injury (n = 13) according to three strategies. Black solid line represents the average electrical impedance tomography (EIT) crossing point between collapse and overdistension. Black dashed line represents the average highest respiratory system compliance (CRS). Black dotted line represents the average end-expiratory transpulmonary pressure (PL) slightly positive. Error bars represent standard error

Full size image
Fig. 5
figure 5

Optimal PEEP and regional respiratory system mechanics in asymmetrical lung injury (n = 13). A The optimal positive end-expiratory pressure (PEEP) according to different PEEP titration strategies in pigs with asymmetrical lung injury: CRS, highest respiratory system compliance; EIT, electrical impedance tomography crossing point; and PL, end-expiratory transpulmonary pressure slightly positive. * represents post-hoc p < 0.05. B Dependent (grey) versus non-dependent (orange) regions. C Left (grey) versus right (orange) regions. Surfactant lavage and high-stretch ventilation were performed in the left side (more injured), while the right side was kept collapsed (less injured). Black solid line represents the average EIT crossing point between collapse and overdistension. Black dashed line represents the average highest CRS. Black dotted line represents the average end-expiratory PL slightly positive. Error bars represent standard error

Full size image

Regarding regional respiratory mechanics in asymmetrical lung injury, right versus left regional CRS and dependent versus non-dependent regional CRS were not significantly different across the three approaches (Fig. 5). The mean absolute difference between ventral and dorsal regional ventilation tended to be higher at PEEP based on end-expiratory PL (38 ± 21%), than at PEEP based on EIT (27 ± 24%, p = 0.080) and at PEEP based on total CRS (26 ± 25%, p = 0.080), Supplemental Fig. 5. The mean absolute difference between right and left percentage of ventilation was 20 ± 15% at PEEP based on end-expiratory PL, 16 ± 12% at PEEP based on EIT, and 14 ± 11% at PEEP based on total CRS, p = 0.568.

2) Impact on pulmonary hemodynamics

In pigs with asymmetrical lung injury, a quadratic relationship was also found between PEEP and mean PAP (Fig. 6A) and PEEP and RV systolic transmural pressure (y = 47.0–0.14x + 7.78x2; R2 = 0.93, p < 0.001). The quadratic model for mean PAP and PEEP (AIC = 11.8) also provided a better fit compared to the linear model (AIC = 47.0). The conditional R2 was 0.811, indicating that 81.1% of the variance in mean PAP can be explained by the model, while the marginal R2 was 0.113, reflecting the variance explained by the fixed effects alone, respectively, with a random effect variance of 40.73 for the intercept and a residual variance of 11.01. The β coefficients for PEEP and PEEP squared in the mixed model analysis were − 0.54 (SE = 0.116, p < 0.001) and 0.044 (SE = 0.006, p < 0.001). In nine pigs of this series of experiments, we measured PVR to verify if the response in mean PAP to PEEP would represent the response in PVR, and we found that PVR behaved like mean PAP (Fig. 6B). Systemic hemodynamics behaved similarly to pigs with bilateral lung injury (Supplemental Fig. 6).

Fig. 6
figure 6

Pulmonary hemodynamics in asymmetrical lung injury (n = 13): A Mean pulmonary artery pressure (Mean PAP) at each level of positive end-expiratory pressure (PEEP). B Pulmonary vascular resistance (PVR) at each level of PEEP. Dots represent mean and error bars represent standard error. Red line represents quadratic regression, and grey shaded area represents 95% confidence interval. Black solid line represents the average electrical impedance tomography (EIT) crossing point between collapse and overdistension. Black dashed line represents the average highest respiratory system compliance (CRS). Black dotted line represents the average end-expiratory transpulmonary pressure (PL) slightly positive

Full size image

Sensitivity analysis

To assess whether changes in PaCO2 were responsible for the observed variability in PVR, we conducted a sensitivity analysis using data from PEEP titrations performed in volume-controlled ventilation (keeping minute ventilation constant), where CO2 levels remained stable. The results of this analysis were consistent with the original findings. In this small subgroup (n = 5), the quadratic relationship between PEEP and mean PAP was also present, with a β coefficients for PEEP of − 0.51 (SE = 0.177, p = 0.005) and for PEEP squared of 0.024 (SE = 0.009, p = 0.006). (Fig. 7A). When adjusting the model by EtCO2 and subject in a mixed model, the effect of EtCO2 was not significant (β coefficient = − 0.023, SE = 0.046, p = 0.617), while PEEP still had a quadratic relationship with mean PAP (β coefficient for PEEP of − 0.62 (SE = 0.20, p = 0.003) and for PEEP squared of 0.027 (SE = 0.0099, p = 0.006). Additionally, PaCO2 was measured at each PEEP level in one experiment, showing stable values (70–80 mm Hg) without a quadratic relationship (R2 = 0.15, p = 0.524). In animals submitted to PEEP titration on pressure-controlled ventilation, when assessing the relationship between PEEP and mean PAP, adjusted by EtCO2 and accounting for inter-case variability, EtCO2 also influenced mean PAP (Supplemental Fig. 7), with a β coefficient of 0.236 (SE = 0.022, p < 0.001), indicating that higher EtCO2 levels are associated with increased mean PAP.

Fig. 7
figure 7

Sensitivity analysis in 5 animals in volume-controlled ventilation. A Mean pulmonary artery pressure (Mean PAP) at each level of positive end-expiratory pressure (PEEP), n = 5. B End-tidal carbon dioxide (EtCO2) at each level of PEEP, n = 5. C Partial pressure of arterial carbon dioxide (PaCO2) at each level of PEEP measured in one of the experiments. Error bars represent standard error. Grey shaded area represents 95% confidence interval

Full size image

Discussion

In porcine models of bilateral and highly recruitable or asymmetrical and less recruitable acute lung injury, we found that: 1) PEEP titration based on EIT, CRS, and end-expiratory PL, resulted in three different levels of optimal PEEP in bilateral lung injury; 2) differences in optimal PEEP were smaller but still significant in asymmetrical lung injury; 3) PEEP had a quadratic relationship with mean PAP (i.e., high mean PAP at low and high levels of PEEP), RV systolic transmural pressure, and PVR; and 4) PEEP set at the EIT crossing-point or highest CRS resulted in more homogeneous regional ventilation than at slightly positive end-expiratory PL in bilateral lung injury.

In our study we measured the “optimal PEEP” based on three different approaches: at the EIT-crossing point, at slightly positive end-expiratory PL, and at the highest CRS. Similarly to previous studies [15, 37, 38], we observed that the three strategies resulted in three different levels of PEEP in our model of bilateral lung injury. Possible explanations for this finding are the different principles and limitations of each strategy and the relative impact of the different parts of the lungs and their distension. Setting PEEP either at the highest CRS or at slightly positive end-expiratory PL aims to achieve the maximal recruitment of the lungs and prevent atelectasis, based on respiratory mechanics [39]. However, global CRS measured by the ventilator is mostly reflecting the dependent part of the lung in bilateral injury. In addition PL provides information about the distending pressure of the lungs more specifically at the level of the esophagus [10, 40,41,42,43]. The EIT, on the other hand, accepts some level of collapse in order to achieve the best compromise between collapse and overdistention [11], estimated by changes in pixel compliance [44].

In the series of experiments with asymmetrical lung injury, we observed low to moderate recruitability, while the bilateral model was highly recruitable [35]. This could explain the lower mean optimal PEEP for all approaches, and a more pronounced increase in overdistention and decrease in CRS while increasing PEEP in the asymmetrical lung injury. Our findings reinforce that the variability in PEEP levels recommended by different approaches extended beyond the global respiratory mechanics, and it highlights the need for a comprehensive understanding of how PEEP impacts other critical physiological aspects, such as regional lung mechanics and pulmonary vascular responses.

For decades, it was widely believed that any increase in PEEP would invariably lead to an increase in PVR due to the expected physiological increase in end-expiratory lung volume, based on the landmark work of Simmons et. al., in 1961 [26], showing that PVR increases with either decrease or increase of lung volume from FRC. In our study, we observed a quadratic relationship (U shape) between PEEP and mean PAP, RV systolic transmural pressure, and PVR. These results support recent evidence [31] that PVR only decreases when PEEP results in lung recruitment. This may help explain previous evidence [45] that ARDS patients have different PVR responses to PEEP. The impairment in pulmonary hemodynamics at high PEEP may be explained by the effect of lung overdistention [26, 27, 46,47,48], compressing the intra-alveolar vessels, while the impairment at low PEEP may be caused not only by stretching extra-alveolar vessels, but also by atelectasis-induced vascular leaking, capillary folding, and hypoxic vasoconstriction [5, 20, 21, 49, 50], as represented in Supplemental Fig. 8. This would explain why we observed a higher increase in mean PAP at low levels of PEEP in pigs with bilateral lung injury (high levels of collapse at low PEEP, recruitable model) and higher increase in mean PAP at high levels of PEEP in pigs with asymmetrical lung injury (high levels of overdistention at high PEEP, poorly recruitable model). To support our hypothesis with EIT the percentage of collapse increased to an average of 60% at low PEEP and overdistention to 30% at high PEEP in the bilateral injury, while overdistention rose to 70% at high PEEP and collapse to 40% at low PEEP in the asymmetrical injury. These findings may explain the lower range of cardiac output and higher CVP observed in the asymmetrical model.

Our results suggest that in mechanically ventilated patients with acute lung injury there might be an optimal range of PEEP for the pulmonary hemodynamics, that results in an end-expiratory lung volume optimal for PVR. This hypothesis has also been suggested by Schmitt et. al. in 2001 [51], based on pulmonary artery Doppler flow velocity measurements in ARDS, comparing three levels of PEEP. All three PEEP selection approaches used in our study tended to recommend a level of PEEP within the lower levels of mean PAP in a decremental PEEP titration, especially EIT. Although the three approaches are grounded in different principles, they all focus on individualizing treatment based on patient-specific respiratory mechanics. This patient-centered strategy ensures that PEEP levels remain within a safe range of respiratory system mechanics, thereby creating favorable conditions for pulmonary hemodynamics.

In terms of regional mechanics, PEEP set based on the EIT and CRS resulted in relative homogenization of regional ventilation in bilateral lung injury, as expected based on their individualized approach. PEEP set at slightly positive end-expiratory PL tended to result in a higher difference between dorsal and regional ventilation. This disparity could be attributed to the fact that esophageal pressure primarily reflects pleural pressure in lung regions adjacent to the esophageal balloon (dependent and middle areas) [52]. As a result, significant variations in regional pleural or transpulmonary pressures might be either overestimated or underestimated when using the direct method to estimate PL [39,40,41], as in our study. Our results highlight that a higher end-expiratory PL should be targeted to keep the lung open (at least 2 cmH2O instead of at least 0 cmH2O), as hypothesized by previous study [53]. In the asymmetrical lung injury, the differences in regional CRS at the optimal PEEP were high regardless of the method of PEEP titration, while in the bilateral injury they were sometimes close to zero (based on EIT and CRS). The regional CRS in the non-dependent region of the lung was higher than in the dependent region, even at higher PEEP. This may be explained by the heterogeneity of injury in the asymmetrical model [32, 33], as only one lung was submitted to surfactant lavage and high stretch (more injured), while the other one was kept collapsed (less injured). This finding is supported by the fact that the regional CRS in the lung that was kept collapsed (right) was always higher than in the other lung (left).

Our study has several limitations. First, one could argue that the two models of lung injury used in this study have underlying differences. However, our primary objective was to compare the impact of PEEP on pulmonary hemodynamics across different lung injury models, rather than to compare the models themselves. The use of various lung injury models was intended to represent different clinical scenarios and provide a comprehensive understanding of the mechanisms involved. Our study design and analysis focused on intra-model comparisons to reduce these potential confounding factors. Secondly, we did not measure PVR for all experiments. However, mean PAP is an adequate surrogate, and we observed a similar relationship between PVR and PEEP in the animals with these measurements. Measuring PVR and cardiac output considerably lengthened the PEEP titration process (with a risk of instability of the model over time), which is the reason that we measured PAP only in most experiments. Thirdly, most PEEP titrations were conducted in pressure-controlled ventilation (varying tidal volume), with consequent variability of PaCO2, which could also influence PVR. To determine if PaCO2 variability drove changes in PVR, we performed a sensitivity analysis using data from volume-controlled ventilation experiments, where CO2 levels were more stable, and the results were consistent. Moreover, it is important to note that the sensitivity analysis was conducted only in a small subsample, which may have influenced the results. Future research could focus on the PaCO2 impact on PVR. Fourthly, it's important to note that all the PEEP titration strategies examined were individualized approaches based on some aspect of global respiratory mechanics. Therefore, we cannot extrapolate our findings to methods based on other variables, such as oxygenation. Finally, we acknowledge that critical illness adds significant complexity to the interaction between PEEP and hemodynamics. Individualized PEEP titration may optimize pulmonary hemodynamics, while lower PEEP levels could offer immediate advantages for systemic hemodynamics. However, these advantages are short-term results, and it is unlikely that keeping a lower PEEP while exposing the right ventricle to a higher afterload can ultimately be beneficial. If the right ventricular function worsens the left ventricle will follow what is supplied by the right side. Our findings provide a foundational understanding of these dynamics and clinicians must also account for factors such as pre-existing cardiovascular dysfunction, fluid balance, and the effects of vasoactive drugs when translating these results into clinical practice.

Conclusions

In porcine models of acute lung injury (highly and poorly recruitable), both very low and very high levels of PEEP may impair pulmonary hemodynamics, increasing mean PAP and potentially PVR either by significantly increasing lung collapse or by causing lung overdistention. The relationship between mean PAP and PEEP follows a “U shape”, with the best range observed at levels of PEEP individualized based on respiratory mechanics, especially by EIT and esophageal pressure monitoring.

Data availability

Data is provided within the manuscript or supplementary information files.

References

  1. Slutsky AS, Ranieri VM. Ventilator-induced lung injury. N Engl J Med. 2013;369:2126–36.

    Article  CAS  PubMed  Google Scholar 

  2. Pham T, Rubenfeld GD. Fifty years of research in ards the epidemiology of acute respiratory distress syndrome a 50th birthday review. Am J Respir Crit Care Med. 2017;195:860–70.

    Article  PubMed  Google Scholar 

  3. Henderson WR, Chen L, Amato MBP, Brochard LJ. Fifty years of research in ARDS respiratory mechanics in acute respiratory distress syndrome. Am J Respir Crit Care Med. 2017;196:822–33.

    Article  PubMed  Google Scholar 

  4. Gattinoni L, Marini JJ. In search of the Holy Grail: identifying the best PEEP in ventilated patients. Intensive Care Med. 2022;48:728–31.

    Article  PubMed  PubMed Central  Google Scholar 

  5. Sousa MLA, Katira BH, Bouch S, Hsing V, Engelberts D, Amato M, et al. Limiting overdistention or collapse when mechanically ventilating injured lungs: a randomized study in a porcine model. Am J Respir Crit Care Med. 2024. https://doi.org/10.1164/rccm.202310-1895OC.

    Article  PubMed  PubMed Central  Google Scholar 

  6. Millington SJ, Cardinal P, Brochard L. Setting and titrating positive end-expiratory pressure. Chest. 2022;161:1566–75.

    Article  PubMed  Google Scholar 

  7. Heunks L, Piquilloud L, Demoule A. How we approach titrating PEEP in patients with acute hypoxemic failure. Crit Care. 2023;27:415.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Depta F, Zdravkovic M, Gentile MA. Should we continue searching for the single best PEEP? Intensive Care Med Exp. 2022;10:9.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Cunningham D, et al. Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome. N Engl J Med. 2004;351:327–36.

    Article  Google Scholar 

  10. Akoumianaki E, Maggiore SM, Valenza F, Bellani G, Jubran A, Loring SH, et al. The application of esophageal pressure measurement in patients with respiratory failure. Am J Respir Crit Care Med. 2014;189:520–31.

    Article  PubMed  Google Scholar 

  11. EL Costa V, Borges JB, Melo A, Suarez-Sipmann F, Toufen C, Bohm SH, et al. Bedside estimation of recruitable alveolar collapse and hyperdistension by electrical impedance tomography. Intensive Care Med. 2009;35:1132–7.

    Article  PubMed  Google Scholar 

  12. Jonkman AH, Alcala GC, Pavlovsky B, Roca O, Spadaro S, Scaramuzzo G, et al. Lung recruitment assessed by electrical impedance tomography (RECRUIT): a multicenter study of COVID-19 ARDS. Am J Respir Crit Care Med. 2023. https://doi.org/10.1164/rccm.202212-2300OC.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Kacmarek RM, Villar J, Sulemanji D, Montiel R, Ferrando C, Blanco J, et al. Open lung approach for the acute respiratory distress syndrome. Crit Care Med. 2016;44:32–42.

    Article  CAS  PubMed  Google Scholar 

  14. Cavalcanti AB, Suzumura ÉA, Laranjeira LN, de Paisani D, Damiani LP, Guimarães HP, et al. Effect of lung recruitment and titrated positive end-expiratory pressure (PEEP) vs low peep on mortality in patients with acute respiratory distress syndrome. JAMA. 2017;318:1335.

    Article  PubMed  PubMed Central  Google Scholar 

  15. Pavlovsky B, Desprez C, Richard J-C, Fage N, Lesimple A, Chean D, et al. Bedside personalized methods based on electrical impedance tomography or respiratory mechanics to set PEEP in ARDS and recruitment-to-inflation ratio: a physiologic study. Ann Intensive Care. 2024;14:1.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Ashbaugh D, Boyd Bigelow D, Petty T, Levine B. Acute respiratory distress in adults. Lancet. 1967;290:319–23.

    Article  Google Scholar 

  17. Suter PM, Fairley HB, Isenberg MD. Optimum end-expiratory airway pressure in patients with acute pulmonary failure. N Engl J Med. 1975;292:284–9.

    Article  CAS  PubMed  Google Scholar 

  18. Dantzker DR, Lynch JP, Weg JG. Depression of cardiac output is a mechanism of shunt reduction in the therapy of acute respiratory failure. Chest. 1980;77:636–42.

    Article  CAS  PubMed  Google Scholar 

  19. Webb HH, Tierney DF. Experimental pulmonary edema due to intermittent positive pressure ventilation with high inflation pressures. Protection by positive end expiratory pressure. Amer Rev Resp Dis. 1974;110:556–65.

    CAS  PubMed  Google Scholar 

  20. Duggan M, McCaul CL, McNamara PJ, Engelberts D, Ackerley C, Kavanagh BP. Atelectasis causes vascular leak and lethal right ventricular failure in uninjured rat lungs. Am J Respir Crit Care Med. 2003;167:1633–40.

    Article  PubMed  Google Scholar 

  21. Tsuchida S, Engelberts D, Peltekova V, Hopkins N, Frndova H, Babyn P, et al. Atelectasis causes alveolar injury in nonatelectatic lung regions. Am J Respir Crit Care Med. 2006;174:279–89.

    Article  PubMed  Google Scholar 

  22. Chiumello D, Carlesso E, Cadringher P, Caironi P, Valenza F, Polli F, et al. Lung stress and strain during mechanical ventilation for acute respiratory distress syndrome. Am J Respir Crit Care Med. 2008;178:346–55.

    Article  PubMed  Google Scholar 

  23. Meade MO, Cook DJ, Guyatt GH, Slutsky AS, Arabi YM, Cooper DJ, et al. Ventilation strategy using low tidal volumes, recruitment maneuvers, and high positive end-expiratory pressure for acute lung injury and acute respiratory distress syndrome. JAMA. 2008;299:637.

    Article  CAS  PubMed  Google Scholar 

  24. Mercat A, Richard J-CM, Vielle B, Jaber S, Osman D, Diehl J-L, et al. Positive end-expiratory pressure setting in adults with acute lung injury and acute respiratory distress syndrome. JAMA. 2008;299:646.

    Article  CAS  PubMed  Google Scholar 

  25. Dianti J, Tisminetzky M, Ferreyro BL, Englesakis M, Del Sorbo L, Sud S, et al. Association of positive end-expiratory pressure and lung recruitment selection strategies with mortality in acute respiratory distress syndrome: a systematic review and network meta-analysis. Am J Respir Crit Care Med. 2022;205:1300–10.

    Article  PubMed  Google Scholar 

  26. Simmons DH, Linde LM, Miller JH, O’Reilly RJ. relation between lung volume and pulmonary vascular resistance. Circ Res. 1961;9:465–71.

    Article  Google Scholar 

  27. Whittenberger JL, McGregor M, Berglund E, Borst HG. Influence of state of inflation of the lung on pulmonary vascular resistance. J Appl Physiol. 1960;15:878–82.

    Article  CAS  PubMed  Google Scholar 

  28. Burton AC, Patel DJ. Effect on pulmonary vascular resistance of inflation of the rabbit lungs. J Appl Physiol. 1958;12:239–46.

    Article  CAS  PubMed  Google Scholar 

  29. Hakim TS, Gilbert E, Camporesi E. Positive end-expiratory pressure increases capillary pressure relative to wedge pressure in the closed and open chest. J Crit Care. 1993;8:80–6.

    Article  CAS  PubMed  Google Scholar 

  30. Koganov Y, Weiss YG, Oppenheim A, Elami A, Pizov R. Positive end-expiratory pressure increases pulmonary venous vascular resistance in patients after coronary artery surgery. Crit Care Med. 1997;25:767–72.

    Article  CAS  PubMed  Google Scholar 

  31. CappioBorlino S, Hagry J, Lai C, Rocca E, Fouqué G, Rosalba D, et al. The effect of on pulmonary vascular resistance depends PEEP on lung recruitability in ARDS patients. Am J Respir Crit Care Med. 2024;3364:0–1.

    Google Scholar 

  32. Pellegrini M, Sousa MLA, Dubo S, Menga LS, Hsing V, Post M, et al. Impact of airway closure and lung collapse on inhaled nitric oxide effect in acute lung injury: an experimental study. Ann Intensive Care. 2024;14:149.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  33. Bastia L, Engelberts D, Osada K, Katira BH, Damiani LF, Yoshida T, et al. Role of positive end-expiratory pressure and regional transpulmonary pressure in asymmetrical lung injury. Am J Respir Crit Care Med. 2021;203:969–76.

    Article  PubMed  Google Scholar 

  34. Katira BH, Osada K, Engelberts D, Bastia L, Damiani LF, Li X, et al. Positive end-expiratory pressure, pleural pressure, and regional compliance during pronation. An experimental study. Am J Respir Crit Care Med. 2021;203:1266–74.

    Article  PubMed  Google Scholar 

  35. Chen L, Del Sorbo L, Grieco DL, Junhasavasdikul D, Rittayamai N, Soliman I, et al. Potential for lung recruitment estimated by the recruitment-to-inflation ratio in acute respiratory distress syndrome. A clinical trial. Am J Respir Crit Care Med. 2020;201:178–87.

    Article  PubMed  Google Scholar 

  36. Kwan WC, Shavelle DM, Laughrun DR. Pulmonary vascular resistance index: getting the units right and why it matters. Clin Cardiol. 2019;42:334–8.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Su P-L, Lin W-C, Ko Y-F, Su P-F, Chen C-W. Positive end-expiratory pressure selection based on best respiratory system compliance or collapse/hyperdistension curves in patients with acute respiratory distress syndrome: lack of correlation with alveolar recruitment. Intensive Care Med. 2018;44:389–91.

    Article  PubMed  Google Scholar 

  38. Perier F, Tuffet S, Maraffi T, Alcala G, Victor M, Haudebourg A-F, et al. Electrical impedance tomography to titrate positive end-expiratory pressure in COVID-19 acute respiratory distress syndrome. Crit Care. 2020;24:678.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Somhorst P, Mousa A, Jonkman AH. Setting positive end-expiratory pressure: the use of esophageal pressure measurements. Curr Opin Crit Care. 2024;30:28–34.

    Article  PubMed  Google Scholar 

  40. Shimatani T, Kyogoku M, Ito Y, Takeuchi M, Khemani RG. Fundamental concepts and the latest evidence for esophageal pressure monitoring. J Intensive Care. 2023;11:22.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Piquilloud L, Beitler JR, Beloncle FM. Monitoring esophageal pressure. Intensive Care Med. 2024;50:953–6.

    Article  PubMed  Google Scholar 

  42. Milic-Emili J, Mead J, Turner JM. Topography of esophageal pressure as a function of posture in man. J Appl Physiol. 1964;19:212–6.

    Article  CAS  PubMed  Google Scholar 

  43. Milic-Emili J, Mead J, Turner JM, Glauser EM. Improved technique for estimating pleural pressure from esophageal balloons. J Appl Physiol. 1964;19:207–11.

    Article  CAS  PubMed  Google Scholar 

  44. Frerichs I, Amato MBP, van Kaam AH, Tingay DG, Zhao Z, Grychtol B, et al. Chest electrical impedance tomography examination, data analysis, terminology, clinical use and recommendations: consensus statement of the translational EIT development study group. Thorax. 2017;72:83–93.

    Article  PubMed  Google Scholar 

  45. Neidhart PP, Suter PM. Changes of right ventricular function with positive end-expiratory pressure (PEEP) in man. Intensive Care Med. 1988;14:471–3.

    Article  PubMed  Google Scholar 

  46. Jozwiak M, Teboul J-L. Heart-Lungs interactions: the basics and clinical implications. Ann Intensive Care. 2024;14:122.

    Article  PubMed  PubMed Central  Google Scholar 

  47. Magder S, Slobod D, Assanangkornchai N. Right ventricular limitation: a tale of two elastances. Am J Respir Crit Care Med. 2023;207:678–92.

    Article  PubMed  Google Scholar 

  48. Marini JJ, Culver BH, Butler J. Mechanical effect of lung distention with positive pressure on cardiac function. Am Rev Respir Dis. 1981;124:382–6.

    CAS  PubMed  Google Scholar 

  49. Lumb AB, Slinger P. Hypoxic pulmonary vasoconstriction. Anesthesiology. 2015;122:932–46.

    Article  CAS  PubMed  Google Scholar 

  50. Bouferrache K, Vieillard-Baron A. Acute respiratory distress syndrome, mechanical ventilation, and right ventricular function. Curr Opin Crit Care. 2011;17:30–5.

    Article  PubMed  Google Scholar 

  51. Schmitt J, Vieillard-Baron A, Augarde R, Prin S, Page B, Jardin F. Positive end-expiratory pressure titration in acute respiratory distress syndrome patients: Impact on right ventricular outflow impedance evaluated by pulmonary artery Doppler flow velocity measurements. Crit Care Med. 2001;29:1154–8.

    Article  CAS  PubMed  Google Scholar 

  52. Yoshida T, Amato MBP, Grieco DL, Chen L, Lima CAS, Roldan R, et al. Esophageal manometry and regional transpulmonary pressure in lung injury. Am J Respir Crit Care Med. 2018;197:1018–26.

    Article  PubMed  Google Scholar 

  53. Sarge T, Baedorf-Kassis E, Banner-Goodspeed V, Novack V, Loring SH, Gong MN, et al. Effect of esophageal pressure–guided positive end-expiratory pressure on survival from acute respiratory distress syndrome: a risk-based and mechanistic reanalysis of the EPVent-2 trial. Am J Respir Crit Care Med. 2021;204:1153–63.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

None.

Funding

This work was supported by grants from Canadian Institute of Health Research (PJT152898 to MPo and PJT156336 to MPo and LB). LSM received financial support from the European Respiratory Society (Long-Term Research Fellowship 2024).

Author information

Authors and Affiliations

  1. Keenan Centre for Biomedical Research, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Unity Health Toronto, Toronto, Canada

    Mayson L. A. Sousa, Luca S. Menga, Annia Schreiber, Mattia Docci, Fernando Vieira & Laurent Brochard

  2. Interdepartmental Division of Critical Care Medicine, University of Toronto, Toronto, Canada

    Mayson L. A. Sousa, Luca S. Menga, Annia Schreiber, Mattia Docci, Fernando Vieira, Ghislaine Douflé & Laurent Brochard

  3. Translational Medicine Program, Research Institute, Hospital for Sick Children, Toronto, Canada

    Mayson L. A. Sousa, Luca S. Menga & Martin Post

  4. Department of Respiratory Therapy, Rady Faculty of Health Sciences, University of Manitoba, 771 McDermot Avenue, Room 338, Winnipeg, Manitoba, R3M 1S1, Canada

    Mayson L. A. Sousa

  5. Pulmonary Division, University of Sao Paulo, Sao Paulo, Brazil

    Eduardo L. V. Costa & Marcelo B. P. Amato

  6. Pediatric Critical Care Medicine, Department of Pediatrics, Washington University in St Louis, St Louis, USA

    Bhushan H. Katira

  7. Anesthesiology and Intensive Care Medicine, Department of Surgical Sciences, Uppsala University, Uppsala, Sweden

    Mariangela Pellegrini

  8. Hedenstierna Laboratory, Department of Surgical Sciences, Uppsala University, Uppsala, Sweden

    Mariangela Pellegrini

  9. Department of Physiotherapy, Universidad de Concepción, Concepción, Chile

    Sebastian Dubo

  10. Institute of Health Policy, Management, and Evaluation, University of Toronto, Toronto, Canada

    Ghislaine Douflé

  11. Department of Anesthesia and Pain Management, Toronto General Hospital, University Health Network, Toronto, Canada

    Ghislaine Douflé

  12. Department of Physiology, University of Toronto, Toronto, Canada

    Martin Post

Authors
  1. Mayson L. A. Sousa
    View author publications

    Search author on:PubMed Google Scholar

  2. Luca S. Menga
    View author publications

    Search author on:PubMed Google Scholar

  3. Annia Schreiber
    View author publications

    Search author on:PubMed Google Scholar

  4. Mattia Docci
    View author publications

    Search author on:PubMed Google Scholar

  5. Fernando Vieira
    View author publications

    Search author on:PubMed Google Scholar

  6. Bhushan H. Katira
    View author publications

    Search author on:PubMed Google Scholar

  7. Mariangela Pellegrini
    View author publications

    Search author on:PubMed Google Scholar

  8. Sebastian Dubo
    View author publications

    Search author on:PubMed Google Scholar

  9. Ghislaine Douflé
    View author publications

    Search author on:PubMed Google Scholar

  10. Eduardo L. V. Costa
    View author publications

    Search author on:PubMed Google Scholar

  11. Martin Post
    View author publications

    Search author on:PubMed Google Scholar

  12. Marcelo B. P. Amato
    View author publications

    Search author on:PubMed Google Scholar

  13. Laurent Brochard
    View author publications

    Search author on:PubMed Google Scholar

Contributions

MLAS, BK, MPe, MPo, MBPA, and LB designed the study; MLAS, LSM, AS, MD, FV, BK, MPe, SD, and GD conducted the experiments; MLAS and LSM performed data analysis; all authors were involved in data interpretation; MLAS, ELVC, MBPA, and LB developed the manuscript; all authors read, revised, and approved the final manuscript.

Corresponding author

Correspondence to Mayson L. A. Sousa.

Ethics declarations

Conflict of interest

The authors declare no competing interests.

Ethical approval and consent to participate

The animal study protocol was approved by the Animal Care Committee of the Peter Gilgan Centre for Research and Learning (PGRLC) at The Hospital for Sick Children – SickKids, reference number 1000058058.

Consent for publication

Not applicable.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Sousa, M.L.A., Menga, L.S., Schreiber, A. et al. Individualized PEEP can improve both pulmonary hemodynamics and lung function in acute lung injury. Crit Care 29, 107 (2025). https://doi.org/10.1186/s13054-025-05325-7

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s13054-025-05325-7

Keywords

Critical Care

ISSN: 1364-8535