Positive end-expiratory pressure in pediatric acute respiratory distress a narrative review

Authors: Lauren Allen, Palen Mallory, Alexandre T. Rotta, Kyle J. Rehder, Andrew G. Miller

Abstract

Background and Objective

Pediatric acute respiratory distress syndrome (PARDS) is associated with substantial morbidity and mortality. Positive end-expiratory pressure (PEEP) can support alveolar recruitment, improve oxygenation, and mitigate ventilator-induced lung injury (VILI). Despite its central role in lung-protective ventilation, evidence to guide PEEP titration in PARDS remains limited. This review summarizes physiologic principles, existing evidence, and bedside strategies for PEEP management in children.

Methods

We performed a narrative review using PubMed from 1967 to 2025, limited to English-language studies. Search terms included “pediatric acute respiratory distress syndrome”, “positive end-expiratory pressure”, “PEEP”, “positive-pressure respiration”, “mechanical ventilation”, “pressure-volume curves”, “electrical impedance tomography”, “esophageal manometry”, “stress index”, and “ventilator-induced lung injury”. Reference lists from key publications and guidelines were also screened.

Key Content and Findings

Among available strategies, utilization of the low PEEP:FiO2 table from the ARDSnet study has the strongest evidence and is simple to apply at the bedside. Other approaches, such as compliance-based maneuvers, oxygenation-guided incremental-decremental maneuvers, stress index, pressure-volume curves, chest imaging, esophageal manometry, and electrical impedance tomography (EIT) lack robust pediatric outcome data. Of these methods for PEEP titration, esophageal manometry and EIT hold significant promise.

Conclusions

Evidence supports maintaining PEEP at or above ARDSNet lower-table recommendations with close attention to physiologic response. Adjunctive monitoring with esophageal manometry or EIT may help balance recruitment and overdistension, but pediatric-specific trials are needed to define best practice. Standardized multicenter studies will be essential to establish evidence-based PEEP strategies in PARDS.

Introduction

Pediatric acute respiratory distress syndrome (PARDS) is a life-threatening condition comprising approximately 3% of all pediatric intensive care unit (PICU) admissions (1,2). Mortality rates in epidemiologic studies range from 10–15% for mild and moderate PARDS, to 33% for severe PARDS, though some clinical trials have reported mortality rates greater than 50% (1-3). Given the significant morbidity and mortality from PARDS, understanding best practices for ventilator management is essential to reduce the risk of ventilator-induced lung injury (VILI). Mortality in PARDS has been associated with supraphysiologic tidal volume (VT) (4), elevated driving pressure (5,6), high mechanical power (7-9), and insufficient application of positive end-expiratory pressure (PEEP) (10). Among these parameters, consensus on optimal PEEP levels in PARDS remains elusive. This narrative review examines the role of PEEP in PARDS, emphasizing its physiologic basis, available clinical evidence, and practical bedside application, while outlining key gaps that warrant further study. We present this article in accordance with the Narrative Review reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-aw-684/rc).

Methods

We conducted a search using PubMed to identify studies relevant to PEEP in PARDS. Table 1 summarizes the search strategy. The search was performed in July 2025 and covered publications from 1967, when acute respiratory distress syndrome (ARDS) was first described, through the search date. Both Medical Subject Headings (MeSH) and free-text terms were used, including “pediatric acute respiratory distress syndrome”, “positive end-expiratory pressure”, “positive-pressure respiration”, “PEEP”, “mechanical ventilation”, “pressure-volume curves”, “electrical impedance tomography”, “esophageal manometry”, “stress index”, and “ventilator-induced lung injury”. Reference lists of key publications and consensus guidelines were also reviewed to identify additional studies. Only English-language studies were included. Particular attention was given to pediatric studies published after 2015, when the first Pediatric Acute Lung Injury Consensus Conference (PALICC) established standardized PARDS criteria, allowing for more consistent comparisons across pediatric cohorts. As a narrative review, our search was not intended to be exhaustive, but rather to capture physiologic and clinical evidence most relevant to bedside PEEP management in PARDS.

PARDS and PALICC

ARDS was first formally recognized in 1967, when Ashbaugh and colleagues described a syndrome of acute-onset hypoxemia, tachypnea, and decreased lung compliance, typically arising in the context of precipitating insults such as pneumonia, trauma, or sepsis (11). Nearly five decades later, the 2015 Pediatric Acute Lung Injury Consensus Conference (PALICC) established the first pediatric-specific definition, introducing the term PARDS.

PALICC defined PARDS as non-cardiogenic pulmonary edema with new radiographic infiltrates and hypoxemia requiring invasive or noninvasive mechanical ventilation (12). Patients must be younger than 18 years, have symptoms within 7 days of a known insult, and lack alternative explanations such as cardiac failure or fluid overload. Severity is stratified using oxygenation index (OI) or oxygen saturation index (OSI), with OI ≥4 or OSI ≥5 required for diagnosis and progressively higher values indicating increasing severity. For patients on noninvasive support, partial pressure of oxygen (PaO2)/fraction of inspired oxygen (FiO2) or pulse oximetry (SpO2)/FiO2 ratios are used in a similar manner. Children with cyanotic heart disease or chronic lung disease are assessed using the same framework, with oxygenation interpreted relative to baseline.

A large, multicenter, multinational point prevalence study noted a 3.2% PARDS prevalence among all PICU admissions with an overall mortality rate of 17.2% (2). Guidelines for the diagnosis and management of PARDS were updated in 2023 through a more inclusive and expanded consensus conference (PALICC-2) (13). This standardized definition of PARDS created a common language among clinicians and researchers in the field, facilitating consistent enrollment criteria for studies and addressing the substantial shortcomings of using adult-centric ARDS criteria in pediatric patients.

Physiology and rationale for PEEP

In healthy lungs, spontaneous breathing occurs through negative intrathoracic pressure generation as the diaphragm contracts (flattens) and the chest wall expands. The resulting pressure gradient draws air through the airways into the alveoli for gas exchange. During normal exhalation, diaphragmatic relaxation and chest wall recoil enable passive exhalation. The volume of gas remaining in the lungs after a normal tidal exhalation is the functional residual capacity (FRC) (14).

At normal FRC, the inward elastic recoil of the lungs is balanced by the outward recoil of the chest wall, placing the respiratory system on the steepest portion of its pressure-volume relationship and at a point of maximal compliance (15,16). At this point, small changes in transpulmonary pressure generate relatively large changes in volume, minimizing the elastic work required to initiate the subsequent breath. This balance of forces is further supported by resistance to expiratory airflow imposed by the upper airway, particularly at the level of the larynx. This so-called “laryngeal braking” is especially relevant in infants with highly compliant chest walls, where glottic narrowing during expiration helps preserve end-expiratory lung volume (EELV) and prevent airway collapse (17). Lung volume maintenance is also aided by surfactant secreted by type II pneumocytes, which lowers alveolar surface tension and prevents collapse at low lung volumes (18). In addition, the alveolar-capillary barrier further contributes to alveolar integrity and maintenance of EELV by limiting fluid transudation into the alveolar space (19).

PARDS disrupts these protective mechanisms that preserve alveolar patency and lung compliance through direct and indirect injury (19). Surfactant production becomes impaired and existing surfactant is inactivated by inflammatory mediators, cellular debris, and proteinaceous edema fluid that flood the alveoli. This results in increased alveolar surface tension and decreased compliance that leads to alveolar collapse. Neutrophil-mediated injury to the alveolar-capillary membrane increases permeability, allowing protein-rich fluid into the alveolar space (19). The resulting lung injury is heterogeneous, as consolidation, edema, and atelectasis typically involve the dependent lung, while the non-dependent lung regions remain relatively preserved (20).

VILI arises from multiple mechanisms, including shearing forces from repetitive opening and closing of alveoli (atelectrauma), regional overdistension with excessive strain (volutrauma), and elevated stress (barotrauma) (21). Lung-protective ventilation strategies minimize strain by employing physiologic VT targets (4–8 mL/kg), and reduce lung stress by limiting plateau pressure (≤28–30 cmH2O) and driving pressure (≤15 cmH2O) (13). Regular assessment of lung-protective ventilation parameters is strongly recommended as part of a VILI-avoidant strategy (22).

The application of PEEP complements these lung-protective strategies in the intubated patient with PARDS by maintaining alveolar recruitment and preventing end-expiratory collapse (Figure 1) (1,23). In PARDS, EELV is already reduced due to surfactant dysfunction, alveolar flooding, and heterogeneous lung injury. Following intubation, sedation-related reductions in respiratory muscle tone, supine positioning, and bypass of the upper airway by the endotracheal tube further predispose the lung to loss of FRC and airway closure at end expiration (15,16). As a result, the respiratory system operates at lower lung volumes, increasing susceptibility to atelectasis and cyclic alveolar derecruitment.

In this setting, appropriately titrated PEEP stabilizes lung volumes by maintaining FRC, thereby preventing airway closure and repetitive collapse of unstable lung units (24). Additional benefits of PEEP include mitigating the loss of the natural laryngeal resistor in infants, increased alveolar surface area, improved ventilation/perfusion (V/Q) matching, and decreased pulmonary vascular resistance (PVR). PVR is reduced by restoring lung volumes to above closing capacity, the lung volume at which small airways begin to collapse (25). Excessive PEEP, however, can result in lung overdistension, increase right ventricular afterload, impair hemodynamics, worsen gas exchange, and paradoxically increase physiologic dead-space by promoting V/Q mismatch (26,27).

Adult literature provides conflicting evidence regarding whether a higher PEEP reduces mortality (28-30). Pediatric data, limited to observational and non-randomized controlled trials (RCTs), support the use of at least moderate PEEP (2,10,31). Despite this evidence, PEEP application varies widely in clinical practice (10,32,33).

Setting and adjusting PEEP

The 2015 PALICC guidelines, developed alongside the original PARDS definition, provided the first consensus recommendations for pediatric PEEP management (12). These initial guidelines, extrapolated primarily from adult data, provided a weak recommendation for the application of moderately elevated levels of PEEP (10–15 cmH2O) titrated to oxygenation response, acknowledging that levels >15 cmH2O may occasionally be necessary. The PALICC-2 2023 update recommended that clinicians employ PEEP levels set at or above those outlined in the ARDSNet lower PEEP:FiO2 table (Table 2) to guide lung-protective ventilation (13). In addition, the current guidelines recommend balancing PEEP benefits and risks through individualized PEEP titration (13).

Across all titration strategies, the physiologic goal of PEEP application is to maintain EELV to reduce cyclic derecruitment and shunt while avoiding regional overdistention and hemodynamic compromise. Because recruitment and overdistension are not directly observable at the bedside, clinicians employ various PEEP titration strategies that rely on surrogate signals to infer changes in lung volume and mechanics. These approaches differ primarily in the physiologic domain they interrogate including oxygenation, global respiratory system mechanics, transpulmonary pressure, or regional ventilation distribution. Consequently, they vary in how directly they reflect recruitment, overdistention, and chest wall effects. As a result, each strategy has inherent limitations when its surrogate signal becomes uncoupled from the underlying lung mechanics it is intended to represent. This section reviews the principal approaches used to guide PEEP adjustment in PARDS, organized by the physiologic signal they emphasize, and discusses the rationale, supporting evidence, and practical limitations of each.

PEEP:FiO2 tables

First incorporated in the ARDSNet ARMA trial (4), PEEP:FiO2 tables remain the most widely adopted PEEP titration method, linking PEEP selection to FiO2 requirement. The PEEP:FiO2 table prescribes stepwise 1–2 cmH2O PEEP increases as FiO2 requirement rises (Table 2). For example, patients requiring an FiO2 of 0.8 receive 14 cmH2O of PEEP. Although the ARMA study the table originates from was primarily designed to test the effect of two VT strategies (12* vs. *6 mL/kg), a PEEP:FiO2 table was included in the protocol to standardize ventilator management across centers, reduce clinician variability, and minimize the risks of prolonged exposure to high FiO2.

The PEEP:FiO2 table is based on the premise that patients with more severe lung injury, as reflected by higher FiO2 requirements, benefit from higher PEEP levels. The physiologic rationale underlying this approach is that escalating FiO2 requirements are assumed to primarily reflect increasing intrapulmonary shunt due to alveolar collapse and loss of FRC. Therefore, higher PEEP is expected to recruit atelectatic lung units and improve V/Q matching. However, this assumption may not hold in all patients, as it implicitly assumes recruitability and does not account for interpatient variability in lung and chest wall mechanics. FiO2 requirement does not distinguish recruitable atelectasis from non-recruitable consolidation, diffusion limitation, altered pulmonary blood flow, or other contributors to hypoxemia.

The original ARDSNet table is now commonly referred to as the “low PEEP/FiO2 table”. This designation reflects comparisons with the more aggressive “high PEEP/FiO2 table”, evaluated in a subsequent trial that showed no advantage over its lower counterpart (Table 2) (34). PEEP:FiO2 tables remain widely used because of their simplicity and ease of bedside implementation; however, they presuppose that FiO2 requirements directly reflect lung recruitability, neglect lung mechanics, have uncertain applicability in pediatric populations, and impose rigid prescriptions that constrain individualized care.

No RCTs have evaluated PEEP:FiO2 tables in children with PARDS. A 2018 multicenter retrospective study by Khemani* et al. *analyzed PEEP application in 1,134 patients managed without formalized PEEP protocols (10). Within 24 hours of PARDS diagnosis, 26% received less PEEP than recommended by the ARDSNet protocol, with this below-table group experiencing 2.1 folds higher ICU mortality. PEEP was commonly increased to 10–12 cmH2O but rarely exceeded this threshold, resulting in PEEP values well below table recommendations once FiO2 exceeded 0.60. These findings strongly informed the PALICC-2 recommendation to maintain PEEP levels at or above the aforementioned lower PEEP:FiO2 table (13). A recent multicenter, non-randomized study of lung protective ventilation in children similarly found that adherence to an ARDSNet-like PEEP:FiO2 table was associated with lower mortality (31).

Adult studies testing high and low PEEP:FiO2 tables have shown conflicting results. Three systematic reviews and meta-analyses found no benefit in unselected patients with ARDS (29,30,35,36); however, a network meta-analysis demonstrated significant mortality benefit [relative risk 0.77, 95% confidence interval (CI): 0.60–0.96] for high PEEP without recruitment maneuvers (28). Notably, this meta-analysis included both high versus low PEEP studies and high PEEP with aggressive recruitment maneuvers, a strategy associated with higher mortality in the ART trial (37). The meta-analysis by Briel* et al. *found that patients with more severe ARDS treated with higher PEEP were more likely to survive (36). A positive oxygenation response to higher PEEP has also been associated with reduced mortality (38). In PARDS, available data do not support preferential use of high versus low tables, however existing data highlight the risks of inadequate PEEP application and support adherence to at least the lower-table targets.

Respiratory system mechanics

Mechanics-based approaches use changes in global respiratory system mechanics as a surrogate for recruitment and overdistention. Compliance-based PEEP titration seeks to minimize global lung stress and strain by identifying the PEEP level associated with the lowest driving pressure. In contrast to PEEP:FiO2 tables, which presume lung recruitability based on hypoxemia, compliance-based titration directly assesses how the respiratory system responds to incremental pressure.

PEEP is titrated to the point of optimal compliance (lowest driving pressure) using incremental or decremental maneuvers. During incremental maneuvers, the PEEP is increased by 1–2 cmH2O while monitoring dynamic compliance. In volume-targeted modes, overdistension manifests when the increase in plateau or peak inspiratory pressure exceeds the increase in PEEP. Similarly, in pressure regulated volume control (PRVC), where the ventilator automatically adjusts inspiratory pressure to maintain target VT, overdistension is indicated when the resultant increase in plateau or peak inspiratory pressure exceeds the increase in PEEP. In pressure targeted modes where inspiratory pressure is held constant, overdistension is evidenced by a decrease in delivered VT. At each step, compliance should be recorded with driving pressure obtained using inspiratory and expiratory hold maneuvers. Proper measurement of respiratory system compliance requires patient passivity (e.g., neuromuscular blockade or deep sedation) as any inspiratory or expiratory breathing effort introduces artifact and diminishes accuracy.

The principal limitation of compliance-based titration is that global improvements in compliance may coexist with regional overdistention, and measured respiratory system compliance reflects both lung and chest wall mechanics. Accordingly, changes in compliance may not reliably distinguish beneficial recruitment from injurious overinflation. To our knowledge, no direct evidence supports using compliance-based approaches to set PEEP in PARDS.

The aforementioned ART trial, the largest study of compliance-based PEEP titration in adult ARDS, compared a moderate PEEP strategy with an experimental approach combining aggressive recruitment maneuvers and compliance-based PEEP titration (37). The intervention group experienced higher mortality (hazard ratio 1.2, 95% CI: 1.01–1.42), fewer ventilator-free days (mean difference −1.1, 95% CI: −2.1 to −0.1), greater vasopressor use, and increased barotrauma. While harm likely resulted from aggressive recruitment maneuvers, higher PEEP contribution could not be excluded (day 1 mean PEEP of 16 cmH2O in the intervention group* vs. *12 cmH2O in the control group). The compliance-based PEEP group had higher PaO2:FiO2 and lower driving pressure, but this did not improve outcomes. Given the lack of pediatric data, compliance-based PEEP titration alone cannot be recommended; compliance should instead serve as supplemental data when applying PEEP:FiO2 tables to detect overdistension.

Oxygenation-guided incremental and decremental PEEP maneuvers

Incremental and decremental PEEP maneuvers have also been used to target oxygenation rather than best compliance. In this approach, PEEP is increased incrementally to improve oxygenation, followed by a decremental phase in which PEEP is reduced until oxygenation begins to drop. PEEP is then set just above this inflection point.

Conceptually, this strategy differs from PEEP:FiO2 tables in that it uses dynamic oxygenation response to an individual patient as a surrogate to infer recruitability, rather than presuming recruitability based on static FiO2 requirement. However, both approaches rely on oxygenation as a surrogate marker of lung recruitment and share the limitation that changes in oxygenation may not accurately reflect underlying lung mechanics. Similar to the caveats with PEEP:FiO2 tables, oxygenation-guided titration may fail in patients in whom hypoxemia is driven by mechanisms other than derecruitment, including increased physiologic dead space or high PVR. In these patients, changes in PEEP may have minimal or misleading effects on oxygenation despite substantial alterations in lung volume or overdistension.

This is the approach employed in the ongoing PROSpect trial (39), an international multicenter, randomized 2×2 factorial study designed to determine whether prone positioning and/or high-frequency oscillatory ventilation improve outcomes in children with moderate-to-severe PARDS. No published pediatric data exist evaluating this strategy, and clinicians must also consider SpO2 limitations when employing this method. SpO2 accuracy is affected within low saturation ranges, by poor perfusion, skin pigmentation, and abnormal hemoglobin presence (40-42). To our knowledge, there is no definitive evidence evaluating oxygenation-guided incremental and decremental maneuvers to set PEEP in adults or children with ARDS. At present, oxygenation-guided maneuvers should be considered investigational or adjunctive rather than definitive PEEP-setting strategies in PARDS.

Pressure-volume (P-V) curves

P-V curves have historically been used to guide PEEP setting and adjustment in patients with ARDS. The rationale is that features of the inflation and deflation limbs [e.g., lower inflection point (LIP)] on the static P-V curve may provide insight into alveolar recruitment and derecruitment dynamics (43).

Traditionally, this was accomplished with the “super syringe” technique, in which pre-specified aliquots of gas are delivered incrementally into the airway. After each increment, a pause is imposed to allow equilibration of pressures throughout the respiratory system, thereby creating true static conditions in which resistive forces are not a factor. This method produces a static inflation curve that reflects the intrinsic P-V properties of the respiratory system. The maneuver is then repeated with gas aliquots drawn into the super syringe to create a static deflation curve. To generate a true static P-V curve, the patient must be completely passive to eliminate the influence of respiratory muscle activity (Pmus). Because of its labor-intensive and cumbersome nature, the super syringe technique is used primarily in research settings rather than routine clinical care.

A way to circumvent this issue is to approximate static conditions by performing “quasi-static” maneuvers. These involve the introduction of gas at very low, constant flow, sometimes combined with brief inspiratory holds, to minimize the influence of resistance on the pressure measurements. Some commercial ventilators (e.g., P/V Tool, Hamilton Medical, Bonaduz, Switzerland) can execute such maneuvers, but their utility in children with ARDS has not been systematically evaluated.

Interpretation of the P-V curve has centered on the LIP on the inflation limb, thought to correspond to the pressure at which small airways and collapsed alveoli begin to recruit (Figure 2). However, the LIP is not always present, may vary over time, and should not be regarded as a fixed physiologic threshold. Importantly, alveolar recruitment continues well above the LIP, and setting PEEP 1–2 cmH2O higher provides no assurance that the chosen PEEP is optimal. On the deflation limb, a separate inflection point has been described that may signal the onset of alveolar derecruitment; theoretically, maintaining PEEP above this point could help sustain recruitment. Still, these features are not consistently reliable markers of the “best” PEEP.

A crucial distinction must also be made between static P-V curves and dynamic P-V the latter are the real-time, breath-to-breath tracings routinely displayed on modern ventilators, whereas the former requires the super syringe technique described above. Dynamic loops are generated during ongoing gas delivery and are therefore markedly influenced by airway resistance. Although dynamic loops may resemble static curves in shape, they cannot be used to identify the LIP or the point of derecruitment and provide no valid guidance for PEEP titration; any attempt to do so is misguided. Only static (or quasi-static) curves, obtained under no-flow or very low-flow conditions, can be meaningfully interpreted for this purpose.

Despite these valuable physiologic insights, static P-V curves are cumbersome to obtain, challenging to interpret, time-consuming, and have not been shown to improve outcomes in PARDS. In adults, a small single-center trial found worse outcomes when PEEP was titrated according to P-V curves compared with electrical impedance tomography (EIT) (44). Taken together, the technical complexity, interpretive uncertainty, and lack of pediatric evidence limit the clinical utility of P-V curves for PEEP titration in PARDS, and their routine clinical use or prioritization for further study cannot currently be recommended.

Stress index

The stress index is a bedside tool used to assess alveolar collapse and overdistension based on the pressure-time waveform during volume-controlled ventilation with constant inspiratory flow. It uses within-breath changes in compliance as a surrogate for tidal recruitment or overdistention and its use in PEEP titration has been described in adults with ARDS. The maneuver requires that the patient is passive, placed on volume-controlled ventilation with a constant flow (square flow waveform), while the pressure-time scalar is evaluated (43). In a lung with optimized compliance, airway pressure (Paw) increases linearly throughout inspiration, corresponding to a stress index of 1 (Figure 3). An index >1 indicates overdistension, as compliance decreases near end-inspiration and the curve assumes a concave shape. Conversely, an index <1 suggests tidal recruitment, in which collapsed alveoli reopen early in inspiration resulting in a convex curve. We are not aware of any studies applying the stress index in children, but small physiologic studies in adults noted improvements in respiratory system compliance and lung injury biomarkers (45-47).

Chest imaging

Many centers use chest imaging [radiograph, computed tomography (CT), lung ultrasound] to inform ventilator management, including PEEP adjustment. Imaging-based approaches use structural or aeration changes as indirect surrogates for lung inflation and recruitment. Chest radiographs are commonly used to assess lung inflation (48) and inform PEEP decisions but they have significant limitations. Most PICU chest radiographs are portable anterior-posterior films, which are quite useful for evaluating pulmonary disease and confirming tube or catheter placement, but are poor for accurately assessing lung volumes. Lung volume accuracy depends on factors such as projection angle, exposure, and phase of the respiratory cycle, compounded by the inherent limitations of making volumetric inferences from a two-dimensional image. The timing of portable chest radiographs within the respiratory cycle is usually unknown to the clinical team. Radiographs taken at end-inspiration may appear hyperinflated if VT is excessive, regardless of PEEP. Conversely, radiographs obtained at end-expiration may appear to have low lung volumes. Hyperinflation at end-expiration could be due to excessive PEEP but also gas-trapping from obstructive disease or intrinsic PEEP from mechanical ventilation. A recent study in neonates found diaphragm position on chest radiograph to be poorly correlated with lung volume on CT (49).

Chest CT has been used to guide PEEP adjustment, but primarily in research settings, where it allows precise assessment of lung volumes at different PEEP levels. These investigations have largely involved physiologic studies in adult patients, require transport to the radiology department, and only provide a single time point assessment of an inherently dynamic process. Lung ultrasound has also been used to identify pulmonary injury in ARDS, evaluate severity, and provide real-time assessment of changes in aeration (50,51). This imaging technique is radiation free and can be useful to inform PEEP adjustments or evaluate response to recruitment maneuvers (52) but cannot detect overdistension. We were unable to find any evidence supporting the use of chest radiographs, chest CT, or lung ultrasound to set or adjust PEEP in PARDS.

Technology assisted techniques

Esophageal manometry

Esophageal pressure (Pes) monitoring has also been used to set PEEP through estimation of transpulmonary pressures. Transpulmonary pressure is the difference between Paw and pleural pressure, representing the true distending force applied to the lungs. When using esophageal manometry, a balloon catheter is placed in the distal esophagus near the diaphragm to obtain a surrogate measurement of pleural pressure (Figure 4) (53). Setting PEEP using Pes involves increasing the PEEP until the end-expiratory Pes is 0 to +2 cmH2O, reflecting near zero end-expiratory transpulmonary pressure (Figure 5). Some modern ventilators can display the Pes waveforms directly, though standalone monitors exist. The major advantage of this technique is that it allows for separation between chest wall and lung mechanics, while also measuring respiratory muscle pressure (Pmus) and transpulmonary pressure (54). Disadvantages include the need for technical expertise in catheter placement and interpretation (55), requirement for an additional invasive device with its associated costs, and the inability to detect overdistension.

A recent pediatric clinical trial utilized Pes to manage patients, though not specifically to set PEEP (56). A small single-center retrospective study of 26 pediatric patients (13 managed with Pes) found that Pes use was associated with higher mean airway pressure (mPaw) (18* vs. 15 cmH2O at 24 hours), higher PEEP (12 vs. *9 cmH2O at 24 hours), lower FiO2, higher PaO2/FiO2, and similar oxygenation index over time compared with a group in which PEEP was adjusted in 2 cmH2O increments to achieve best compliance and decrease in FiO2 (57). Surprisingly, there were no differences in plateau pressure, driving pressure, dynamic compliance, or alveolar dead space. Mortality and duration of mechanical ventilation were also similar. The largest adult trial to date found no differences in mortality or ventilator-free days between Pes-guided PEEP and use of the high PEEP:FiO2 table (58). PEEP was similar in both groups over time, as were lung mechanics, PaO2/FiO2, driving pressure, and end-expiratory transpulmonary pressure. A secondary analysis of this trial stratifying patients by an end-expiratory pressure near 0 cmH2O noted improved survival compared to end-expiratory pressure targeted below or above 0 cmH2O (59). The adult literature on esophageal manometry is robust and has been covered in depth elsewhere (53,60). Taken together, these data suggest that Pes monitoring offers no clear advantages over PEEP:FiO2 tables for PEEP titration, though direct comparisons between the two strategies are lacking in PARDS.

EIT

EIT provides radiation-free, real-time dynamic bedside lung volume assessment (61). A flexible belt with surface electrodes encircles the chest, delivering and sensing very low amplitude alternating current. The change in signal between two electrodes, known as the change in thoracic impedance, is measured and reconstructed into cross-sectional images. These images allow real-time assessment of changes in lung volume, the extent of overdistension, and ventilation heterogeneity throughout the lung during PEEP adjustments (Figure 6) (62). Some systems include software capable of performing incremental-decremental PEEP maneuvers that integrate EIT data with lung mechanics to identify a “best PEEP” that balances lung recruitment maintenance against overdistension. Limitations include cost, a steep learning curve, and limited availability. The FDA only recently approved EIT for clinical use in the United States. Consequently, most available data come from research studies, and its use in daily clinical practice remains to be defined.

Limited pediatric data suggest the feasibility and potential benefits of EIT-guided PEEP titration in children. A proof-of-concept study of eight children with early PARDS showed EIT-guided decremental PEEP reduced regional collapse without increasing overdistension, improving compliance and oxygenation using ARDSNet tables as safety guardrails (63). A single-center feasibility cohort study found that EIT-derived “best PEEP” was on average slightly below both the clinically set PEEP and the low PEEP:FiO2 table (64). Conversely, a single-center study of 12 children with PARDS found that PEEP chosen by global* vs. *EIT-based regional compliance was identical in half the cases and differed by only ±2 cmH2O in the remainder, suggesting global compliance may suffice where EIT is unavailable (65).

In adults, a recent systematic review and meta-analysis on EIT for PEEP titration noted significant limitations from study variability and lack of adequate RCTs (66). Within these limitations, EIT-guided PEEP titration often diverged from both the low and high PEEP:FiO2 tables, typically yielding values in-between the two. EIT guidance was associated with modest improvements in compliance and a trend toward lower driving pressure, though these effects were inconsistent across studies (66). Pooled analysis suggested improved survival [risk ratio (RR) 1.54, 95% CI: 1.09–2.18] with EIT versus PEEP:FiO2 tables or compliance-based titration, though these results warrant caution given small sample sizes (132 control; 127 EIT-guided) (57).

A retrospective cohort of 75 patients with ARDS from coronavirus disease 2019 (COVID-19) found EIT identified clinically relevant PEEP adjustments for 63% of patients compared to the ARDSNet high PEEP:FiO2 table (67). About half required higher PEEP for recruitment while the remainder benefited from lower PEEP to avoid overdistension, underscoring the value of individualized titration despite unchanged median cohort PEEP. A prospective multicenter study showed recruitability varied widely independent of ARDS severity, with EIT-defined optimal PEEP differing from compliance-based settings in 81% of patients (68). A randomized crossover pilot study of 12 patients reported that EIT-guided titration set PEEP about 2 cmH2O lower than the high PEEP:FiO2 table. Despite this modest decrease, compliance improved (+8 mL/cmH2O) with attendant reductions in driving (3 cmH2O), plateau (5 cmH2O), and peak inspiratory (6 cmH2O) pressures, resulting in a decrease in mechanical power of ~4 J/min.

Available data suggest that EIT can individualize PEEP in PARDS, potentially reducing both alveolar collapse and overdistension. The absence of large pediatric or adult outcome trials limits the applicability of EIT to daily clinical practice, as do expense and limited availability. Although promising, it remains uncertain whether individualized, EIT-guided PEEP setting improves patient outcomes.

Special situations

Extracorporeal membrane oxygenation (ECMO)

PEEP management in patients receiving ECMO presents unique challenges as oxygenation-based strategies (e.g., PEEP:FiO2 tables, incremental/decremental titration), lack validation due to ECMO circuit-provided oxygenation support. Clinicians typically set PEEP based on lung mechanics, imaging, oxygenation, expected clinical course, or institutional protocol. PEEP can be increased to enhance native lung participation in gas exchange if oxygenation is below target (69,70). A large multicenter study found that PEEP during ECMO was typically set at 10 cmH2O with little variation over time, without association to outcomes (71). Multiple other studies report PEEP rarely exceeds 10 cmH2O (72-74), with several finding PEEP >10 cmH2O during venous-venous-ECMO was associated with an increase in mortality, likely reflecting confounding by severity rather than causation (73,74). Future studies should evaluate EIT and esophageal balloon monitoring for ECMO PEEP optimization.

Transitioning from high-frequency ventilation

Best practices for setting PEEP when transitioning patients from high-frequency ventilation to conventional mechanical ventilation (CMV) are not well defined, as few studies have addressed transition settings. Patients on high-frequency oscillatory ventilation (HFOV) are usually transitioned to CMV once the mPaw reaches 20–22 cmH2O. Initial PEEP usually targets a mPaw 16–18 cmH2O, which usually requires ≥12 cmH2O PEEP, but is often adjusted based on lung mechanics and oxygenation. Similar strategies apply for high-frequency jet ventilation (HFJV) or high-frequency percussive ventilation. A single-center study of infants receiving HFJV with severe respiratory failure noted patients were transitioned from a mPaw of 13 cmH2O and placed on a PEEP of 8 cmH2O on CMV (75). These transitions involve abrupt lung volume and gas exchange changes, requiring careful consideration of individual mechanics and rapid PEEP strategy adaptation to avoid derecruitment or overdistension.

Increased chest wall elastance or abnormal chest wall mechanics

Increased chest wall elastance from volume overload, obesity, or restrictive chest wall disorders presents unique challenges. PEEP may need to be increased to counteract chest wall elastance, though differentiating chest wall from lung mechanics requires esophageal manometry (54). PEEP below the threshold to counteract the chest wall results in atelectasis, higher driving pressure (worse compliance), and poor gas exchange. PALICC guidelines allow plateau pressures up to 32 cmH2O in cases of increased chest wall elastance; however, severe obesity may require higher levels. Clinicians should closely monitor hemodynamics, compliance, and oxygenation response, and consider esophageal manometry in morbidly obese patients.

Discussion and future directions

Although considerable uncertainty and ample knowledge gaps remain regarding optimal PEEP management in PARDS, both adult and pediatric evidence generally support the use of higher PEEP levels in patients requiring higher FiO2. However, studies have consistently demonstrated PEEP underutilization even in the setting of elevated FiO2 requirements (10,31). This PEEP underutilization may reflect clinician hesitance and a lack of standardized PEEP titration protocols. Clinician hesitance to escalate PEEP may stem from concerns about potential adverse effects including localized overdistension and hemodynamic compromise. Limited pediatric data may affect clinician reluctance to increase PEEP, as infants and small children may be perceived as less tolerant of high PEEP than older children or adults. The heterogeneous nature of PARDS may also limit adoption of PEEP:FiO2 tables as individual clinicians may believe they can better personalize PEEP. This underscores the importance of clinicians mastering individual PEEP titration strategies, as appropriate application of PEEP is crucial for optimizing outcomes while avoiding potential complications.

This review is limited by the small body of pediatric evidence and the lack of comparative studies evaluating PEEP:FiO2 tables against alternative strategies. In addition, as a narrative review it is subject to selection bias and was not designed to be exhaustive. Despite these constraints, it synthesizes both pediatric and adult data to provide practical guidance for clinicians while identifying key gaps to be addressed in future research.

Future studies should further investigate the role of EIT and Pes monitoring, leveraging both single-center and multicenter databases. Clinical trials should employ the low PEEP:FiO2 table as control, based on currently available evidence. Physiologic studies should stratify analyses by PARDS severity as differing severities may respond differently to PEEP. Future trials, including small physiologic studies, should employ a master protocol to enable result pooling in subsequent analyses (3). Given the inherent challenges of enrolling children in clinical trials and relative PARDS rarity, standardized data collection and enrollment criteria are critical for generating sufficient sample sizes to evaluate patient-relevant outcomes including mortality, ventilator-free days, and hospital length of stay.

Conclusions

Various methods exist for setting and adjusting PEEP in PARDS, from simple oxygenation-based tables to advanced physiologic monitoring techniques. Current pediatric evidence supports maintaining PEEP at or above the low PEEP:FiO2 table derived from ARDSNet, as PEEP underutilization remains common and is associated with worse outcomes. While compliance-based and oxygenation-based titration, imaging, esophageal manometry, and EIT each offer potential advantages, pediatric data remain limited, with no approach demonstrating superior clinical outcomes. Transpulmonary pressure monitoring and EIT show greatest promise for individualized PEEP titration by separating chest wall and lung mechanics and offering bedside imaging of recruitment and overdistension, respectively. Until stronger pediatric evidence is available, clinicians should master both PEEP:FiO2 tables and physiologic adjuncts, applying them judiciously to optimize oxygenation, minimize ventilator-induced lung injury, and improve outcomes in children with PARDS.

Supplementary