Mechanical power in decelerating flow versus square flow ventilation in pediatric ARDS

Abstract

Background:

Mechanical power (MP) is a summary variable quantifying the risk of ventilator-induced lung injury (VILI). The original MP equation was developed using square flow ventilation. However, most children are ventilated using decelerating flow. It is unclear whether MP differs according to mode of flow delivery. We compared MP in children with acute respiratory distress syndrome (ARDS) who received both square and decelerating flow ventilation.

METHODS:

This was a secondary analysis of a prospectively enrolled cohort of pediatric ARDS. Patients were ventilated on decelerating flow, and then placed in square flow and allowed to stabilize. Ventilator metrics from both modes were collected within 24 hours of ARDS onset. Paired t-tests were used to compare differences in MP between the modes.

RESULTS:

We enrolled 185 subjects with a median oxygenation index of 9.5 (IQR 7, 13) and median age 8.3 years (IQR 1.8, 14). MP was lower in square flow (mean 0.46 J·min−1 Kg−1, SD 0.25, 95% CI 0.42–0.50) than in decelerating flow modes (mean 0.49 J·min−1 Kg−1, SD 0.28, 95% CI 0.45–0.53) with a mean difference of 0.03 J·min−1 Kg−1 (SD 0.08, 95% CI 0.014–0.038) (p<0.001). This result remained statistically significant when stratified by age < 2 years in square flow compared to decelerating flow and also when stratified by age >/= 2 years in square flow compared to decelerating flow. The elastic contribution in square flow was 70% and the resistive contribution was 30%.

CONCLUSIONS:

MP was marginally lower in square flow than in decelerating flow, although the clinical significance of this is unclear. Upward of 30% of MP may go towards overcoming resistance, regardless of age. This is nearly three-fold greater resistance compared to what has been reported in adults.

Summary Statement:

Mechanical power (MP) was marginally lower in square flow than decelerating flow. Upward of 30% of MP may go towards overcoming resistance, regardless of age. This is nearly three-fold greater resistance compared to what has been reported in adults.

Introduction

Ventilator management is highly varied for children with acute respiratory distress syndrome (ARDS), in part, due to the absence of multicenter clinical trials evaluating different ventilation strategies¹. Mechanical power (MP) is a composite variable describing the energy transfer from the ventilator to the respiratory system within a given timeframe². The MP equation incorporates potentially modifiable ventilator parameters that may contribute to ventilator-induced lung injury (VILI) into a single unifying variable. M concept as MP provides a comprehensive measure of the energy load delivered to the lungs, potentially helping to predict and mitigate the risk of VILI and thereby improve patient safety and outcomes. High MP has been associated with adverse outcomes in adults and pediatric ARDS^3–6. In adults, high MP has been associated with lung injury and mortality in both the critical care and intraoperative settings. This association has also been demonstrated in adult patients without ARDS. For example, a retrospective analysis of over 230,000 adults who had elective noncardiac surgery with general anesthesia and mechanical ventilation between 2008 and 2018 found that higher MP during anesthesia was associated with postoperative respiratory failure requiring reintubation⁷. In pediatric studies, high MP has been associated with increased mortality⁶ and fewer ventilator-free days (VFDs) in children with ARDS, with one study showing a stronger association in children under two years old⁵. Initially described in adults, the original MP equation focused on square flow ventilation strategies, such as volume control, with readily separable resistive and elastic components (owing to the ability to attain distinct peak and plateau pressures)². However, most children are ventilated using decelerating flow, such as pressure control or hybrid modes, which do not easily lend themselves to separating resistive and elastic components⁸. It is unclear whether MP differs according to mode of flow delivery. It is additionally unclear, if there were a difference, whether this difference is driven by younger patients who potentially have higher airway resistance and higher incidence of concurrent obstructive processes (e.g., airway plugging) than older children.

Peak airway pressure (PIP) above plateau pressure in a ventilator is proportional to airway resistance and inspiratory flow rate. However, pediatric ARDS, unlike adult ARDS, has a resistive component that may not be well appreciated given the predominant use of decelerating flow ventilation in children. Indeed, in the multinational Pediatric ARDS Incidence and Epidemiology (PARDIE) point-prevalence study, plateau pressures were recorded in < 3% of subjects⁸. However, it is unclear whether the preference favoring decelerating flow modes impacts delivery of MP to children.

Therefore, we performed a prospective cohort study of pediatric ARDS to evaluate differences in MP between decelerating flow and square flow ventilation. We additionally sought to determine the proportion of pressure contributing to resistive and elastic work in square flow ventilation.

Materials and Methods

Study Design and Setting

This was a prospective cohort of pediatric ARDS enrolled between December 2019 and February 2023 from the Children’s Hospital of Philadelphia (CHOP) PICU, approved by the Institutional Review Board (IRB 20–017935) with requirement for informed consent waived. Procedures were followed in accordance with the ethical standards of the CHOP IRB and with the Helsinki Declaration of 1975. Ventilator management was not protocolized; however, all subjects were ventilated on a Drager Evita V500 and initially placed in decelerating flow modes at the provider’s discretion (either pressure-regulated volume control [PRVC] or pressure control ventilation [PCV]). All patients were either under neuromuscular blockade or were not breathing above the set rate. All patients were supine at the time of assessment. Our PICU uses cuffed endotracheal tubes, with all leaks < 15%. Patients were then placed in square flow volume control ventilation (VCV) and allowed to stabilize for 30–60 seconds before conducting a 0.5 second inspiratory pause. Peak inspiratory pressure (PIP) and plateau pressure (Pplat) measurements were then collected in square flow VCV.

Participants and Outcomes

A portion of this cohort has been previously described⁹. In brief, patients were screened daily for study eligibility from December 20, 2019 to February 16, 2023 in the PICU. Intubated children meeting both Berlin¹⁰ and Pediatric Acute Lung Injury Consensus Conferences (PALICC)¹¹ criteria for ARDS (oxygenation index ≥ 4 or oxygenation saturation index ≥ 5 sustained for 4 hours) with bilateral infiltrates on chest radiograph and on conventional ventilation were included. All endotracheal tubes used in this cohort were cuffed with a tube leak less than 15% (95% of patients had a leak less than 5%)¹². Subjects on non-conventional ventilator modes and extracorporeal support at ARDS onset, and those unable to be approached within 24 hours of ARDS onset, were excluded.

Variables and Definitions

Demographics, comorbidities, and ventilator settings were obtained from a single timepoint collected within 24 hours of ARDS onset. Multiple organ dysfunction was quantified using Pediatric Logistic Organ Dysfunction (PELOD-2) score¹³. Shock was quantified using vasopressor score¹⁴. The designation “Immunocompromised” required an immunocompromising diagnosis (oncologic, immunologic, rheumatologic, or transplant) and active immunosuppressive chemotherapy, or a congenital immunodeficiency. MP was calculated from variables collected within 24 hours of ARDS onset using the simplified equations proposed for square and decelerating flow ventilation, normalized to actual body weight. Height was not universally collected, precluding calculation of predicted body weight in all subjects. Additional variables related to VILI, including MP, were calculated from derivations of commonly accepted equations (Table 1). Measurements for resistance and elastance are provided in Table 1. It is important to note that decelerating flow ventilation does not readily allow for the measurement of resistance, as the difference between PIP and plateau pressure is a useful property of square flow ventilation that facilitates these measurements. This fundamental difference underlies the rationale for our study. In the context of the mechanical power equation, resistance and elastance were deconstructed to their separate components. Specifically, resistance was calculated as the difference between PIP and the plateau pressure. Elastance was the sum of static elastance (represented by PEEP) and dynamic elastance (represented by the driving pressure).

Table 1 –

Definition of variables pertaining to ventilator-induced lung injury (VILI).

	Square-Flow	Decelerating-Flow
Mechanical Power (J min−1 kg−1)	[(VT/kg)/1000] × (PIP – [(Plateau minus PEEP)/2] × RR × 0.098	[(VT/kg)/1000] × (PEEP + [PIP minus PEEP]) × RR × 0.098
Driving Pressure (cmH2O)	Plateau minus PEEP	PIP minus PEEP
Mechanical Energy) (J kg −1 )	[(VT/kg)/1000] × (PIP – [(Plateau minus PEEP)/2] × 0.098	VT/kg × (PEEP + [PIP minus PEEP]) × 0.098
Resistive Component of MP	PIP minus Plateau	N/A
Static Elastic Component of MP	PEEP	N/A
Dynamic Elastic Component of MP	0.5 * (Plateau minus PEEP)	N/A

MP (Mechanical power), PIP (Peak inspiratory pressure), Plateau (Plateau pressure), PEEP (Positive end-expiratory pressure), V_T (Tidal volume), RR (Respiratory rate)

Statistical Analysis

We required 40 subjects to detect a 10% difference in MP between ventilator modes. However, our secondary goal was to fully characterize MP in pediatric ARDS, including providing proportions of resistance and elastic power, so we analyzed a larger sample size to be representative. Analyses were performed using Stata 17.0 (StataCorp, College Station, TX). Descriptive summaries were used to evaluate demographic information. Data were non-normally distributed (Shapiro-Wilk test p<0.05). However, given our adequately large sample size and our desire to readily quantify the magnitude of difference between ventilator modes, we analyzed data using parametric tests. Notably, our conclusions were identical when analyzed using non-parametric testing methods. Therefore, paired t-tests were used to compare differences in MP (as well as other individual ventilator components) between square and decelerating flow modes across all ages and stratified by age < 2 years or age >/= 2 years. Paired t-tests were also used to compare differences in MP within each ventilation mode based on age < 2 years or age >/= 2 years. Finally, paired t-tests were used to compare MP in square and decelerating flows based on use of continuous neuromuscular blocking (NMB) agents. In an effort to reduce bias, analyses were performed blinded to clinical data and performed by an investigator not involved in collecting ventilator data. Patients with missing data due to inability to be approached within 24 hours of ARDS onset were excluded from the analysis.

Results

Description of the Cohort

Over the study period, 230 patients had ARDS (Figure 1), of whom 185 were eligible. Of these 185, 148 (80%) were ventilated on pressure-regulated volume control and 37 (20%) on pressure control, and 98 (53%) were under NMB during ventilator measurements. There were 31 (17%) non-survivors (Table 2). Pneumonia (51%) and non-pulmonary sepsis (17%) were the most common causes of ARDS.

Fig. 1 –

Enrollment flowchart.

Table 2 –

Demographics of the cohort. Data are reported as N (%) or median (Interquartile Range).

Variables	All Patients (n=185)
Age, yr	8.3 (1.8–14)
Female (%)	89 (48)
Severity of illness
PELOD-2 score	6 (5–9)
Vasopressor score	4 (0–13)
Immunocompromised (%)	48 (26)
Stem cell transplant (%)	28 (15)
Causes of ARDS (%)
Pneumonia	95 (51)
Aspiration	20 (11)
COVID	13 (7)
Sepsis	32 (17)
Trauma	3 (2)
Chemotherapy/CART therapy	4 (2)
Other	18 (10)
At ARDS onset
PRVC	148 (80)
PCV	37 (20)
FiO2	0.40 (0.35–0.55)
OI (n=110)	9.5 (7–13)
OSI (n=75)	7.2 (6–9.6)
Tidal volume	6.4 (5.5–7.5)
Tidal volume/PBW (n=159)	7.1 (6.2–8.3)
PEEP (cm H2O)	10 (10–12)
Respiratory rate	24 (20–30)
PRVC or PCV (n=185)
PIP (cm H2O)	28.5 (25.2–32.8)
ΔP (PIP minus PEEP) (cm H2O)	18.4 (14.8–21.6)
VCV
PIP (cm H2O)	36.4 (31.3–41.3)
Plateau Pressure (cm H2O)	27.5 (24.2–32)
ΔP (Plateau minus PEEP) (cm H2O)	17.6 (13.7–21)
Ancillary therapies in first 72h (%)
Inhaled nitric oxide	65 (35)
Steroids	95 (51)
Methylprednisone or Prednisone	42 (44)
Hydrocortisone	33 (35)
Dexamethasone	22 (23)
Neuromuscular blockade	98 (53)
Prone positioning	19 (10)
Outcome
Died in PICU before 90 days	31 (17)
Discharged alive from PICU	143 (79)
Remained in PICU > 90 days and alive	7 (4)

PELOD (Pediatric logistic organ dysfunction), ARDS (Acute respiratory distress syndrome), PRVC (Pressure-regulated volume control), PCV (Pressure-controlled ventilation), OI (Oxygenation index), OSI (Oxygen saturation index), PBW (Predicted body weight), PEEP (Positive end-expiratory pressure), PIP (Peak inspiratory pressure), VCV (volume-controlled ventilation)

Mechanical Power in Square Versus Decelerating Flow Ventilation

MP was marginally, but statistically significantly, lower in square flow (mean 0.46 J·min−1 Kg−1, SD 0.25, 95% CI 0.42–0.50) than in decelerating flow modes (mean 0.49 J·min−1 Kg−1, SD 0.28, 95% CI 0.45–0.53) with a mean difference of 0.03 J·min−1 Kg−1 (SD 0.08, 95% CI 0.014–0.038) (p<0.001). (Table 3) (Figure 2). MP between the two different ventilation strategies remained statistically significant when stratified by age. In children < 2 years, MP was lower in square flow (mean 0.64 J·min−1 kg−1, SD 0.24, 95% CI 0.57–0.71) compared to decelerating flow (mean 0.72 J·min−1 kg−1, SD 0.31, 95% CI 0.63–0.81) with a mean difference of 0.07 J·min−1 Kg−1 (SD 0.12, 95% CI 0.04–0.11) (p=0.001). In children >/= 2 years, MP in square flow was lower (mean 0.40 J·min−1 kg−1, SD 0.22, 95% CI 0.36–0.44) compared to decelerating flow (mean 0.41 J·min−1 kg−1 SD 0.21, 95% CI 0.37–0.44) with a mean difference of 0.01 J·min−1 kg−1 (SD 0.05, 95% CI 0.36–0.44) (p=0.04) (Table 4). Notably, within the same ventilation strategy, children < 2 years in square flow had higher MP (mean 0.64 J·min−1 kg−1, SD 0.24, 95% CI 0.57–0.71) than children >/= 2 years (mean 0.40 J·min−1 kg−1, SD 0.22, 95% CI 0.36–0.44) with a mean difference of 0.25 J·min−1 kg−1 (95% CI 0.17–0.32) (p<0.001). This was observed in children exposed to decelerating flow when dichotomized by age < 2 years (mean 0.72 J·min−1 kg−1, SD 0.31, 95% CI 0.63–0.81) and >/= 2 years (mean 0.41 J·min−1 kg−1, SD 0.21, 95% CI 0.37–0.44) with a mean difference of 0.31 J·min−1 kg−1 (95% CI 0.23–0.39) (p<0.001).

Table 3 –

Paired t-test results comparing differences in mechanical properties in decelerating-flow (PCV, PRVC) versus square-flow (VCV) ventilation.

	PCV, PRVC Mean (SD, 95% CI)	VCV Mean (SD, 95% CI)	Effect Size (SD, 95% CI)	p
Mechanical Power	0.49 (0.28, 0.45–0.53)	0.46 (0.25, 0.42–0.50)	0.03 (0.08, 0.014–0.038)	<0.001
PIP	29.4 (6.0, 28.5–30.2)	37.0 (7.9, 35.9–38.2)	7.7 (3.6, 7.1–8.2)	<0.001
Driving Pressure	18.8 (5.3, 18.0–19.5)	17.9 (5.2, 17.1–18.6)	0.89 (0.78, 0.78–1.0)	<0.001
Mechanical Energy (remove RR)	0.019 (0.005, 0.018–0.019)	0.018 (0.005, 0.017–0.018)	0.0008 (0.002, 0.0004–0.001)	<0.001

PCV (Pressure-controlled ventilation), PRVC (Pressure-regulated volume control), IQR (Interquartile range), VCV (volume-controlled ventilation), Mechanical Power (J·min−1 Kg−1), PIP (Peak inspiratory pressure; cmH2O), Driving Pressure (cmH2O), Mechanical Energy (J·Kg−1), RR (Respiratory rate)

Fig. 2 –

Dot plots of mechanical power, mechanical energy, peak inspiratory pressure (PIP), and ΔP (driving pressure) by treatment group (pressure-controlled ventilation versus volume-controlled ventilation). Black line represents group mean.

Table 4 –

Paired t-test results comparing differences in mechanical properties in decelerating-flow (PCV, PRVC) versus square-flow (VCV) ventilation, dichotomized by age.

	PCV, PRVC Mean (SD, 95% CI)	VCV Mean (SD, 95% CI)	Effect size Mean (SD, 95% CI)	p
Mechanical Power
Age < 2 years	0.72 (0.31, 0.63–0.81)	0.64 (0.24, 0.57–0.71)	0.07 (0.12, 0.04–0.11)	0.001
Age >/= 2 years	0.41 (0.21, 0.37–0.44)	0.40 (0.22, 0.36–0.44)	0.01 (0.05, 0.36–0.44)	0.04
PIP
Age < 2 years	28.8 (5.5, 27.2–30.4)	35.3 (7.1, 33.3–37.4)	6.5 (3.6, 5.5–7.5)	<0.001
Age >/= 2 years	29.6 (6.2, 28.5–30.6)	37.6 (8.1, 36.2–39.0)	8.1 (3.5, 7.5–8.7)	<0.001
Driving Pressure
Age < 2 years	18.9 (5.1, 17.4–20.4)	18.1 (5.3, 16.6–19.7)	0.78 (0.70, 0.57–0.98)	<0.001
Age >/= 2 years	18.7 (5.4, 17.8–19.6)	17.8 (5.2, 16.9–18.7)	0.93 (0.80, 0.79–1.1)	<0.001
Mechanical Energy (remove RR)
Age < 2 years	0.021 (0.006, 0.019–0.023)	0.019 (0.005, 0.018–0.021)	0.002 (0.003, 0.001–0.003)	<0.001
Age >/= 2 years	0.018 (0.005, 0.017–0.019)	0.017 (0.005, 0.016–0.018)	0.0004 (0.002, 0.0005–0.0008)	0.03

PCV (Pressure-controlled ventilation), PRVC (Pressure-regulated volume control), IQR (Interquartile range), VCV (volume-controlled ventilation), Mechanical Power (J·min−1 Kg−1), PIP (Peak inspiratory pressure; cmH2O), Driving Pressure (cmH2O), Mechanical Energy (J·Kg−1), RR (Respiratory rate)

Patients placed in square-flow ventilation who received continuous NMB infusions had higher MP (mean 0.50 J·min−1 kg−1, SD 0.22, 95% CI 0.45–0.54) compared to patients who did not receive NMB infusions in square flow ventilation (mean 0.42 J·min−1 kg−1, SD 0.29, 95% CI 0.36–0.48) with a mean difference of 0.07 J·min−1 kg−1 (95% CI 0.001–0.14) that did not reach statistical significance (p=0.0532). Patients placed in decelerating-flow ventilation who received continuous NMB infusions also had higher MP (mean 0.54 J·min−1 kg−1, SD 0.27, 95% CI 0.48–0.59) compared to patients who did not receive NMB infusions (mean 0.43 J·min−1 kg−1, SD 0.28, 95% CI 0.37–0.49) with a mean difference of 0.10 J·min−1 kg−1 (95% CI 0.02–0.18) that did reach statistical significance (p=0.01).

Analysis of Individual Mechanical Properties in Square Versus Decelerating Flow Ventilation

PIP was higher in square flow ventilation (mean 37.0 cmH2O, SD 7.9, 95% CI 35.9–38.2) compared to decelerating flow (mean 29.4 cmH2O, SD 6.0, 95% CI 28.5–30.2) with a mean difference of 7.7 cmH2O (SD 3.6, 95% CI 7.1–8.2) (P<0.001) (Figure 2). Driving pressure was lower in square flow (17.9 cmH2O, SD 5.2, 95% CI 17.1–18.6) compared to decelerating flow 18.8 cmH2O (SD 5.3, 95% CI 18.0–19.5) with a mean difference of 0.89 cmH2O (SD 0.78, 95% CI 0.78–1.0) (P<0.001). Mechanical energy was lower in square flow (median 0.018 J kg⁻¹, SD 0.005, 95% CI 0.017–0.018) compared to decelerating flow ventilation (mean 0.019 J kg⁻¹, SD 0.005, 95% CI 0.018–0.019) with a mean difference of 0.0008 J kg⁻¹ (SD 0.002, 95% CI 0.0004–0.001) (P<0.001). These results all remained significant when stratified by age <2 years and age >/= 2 years. The contributions to mechanical power in square flow ventilation were 30% resistive and 70% elastic (37% static elastic, 33% dynamic elastic) (Figure 3). These proportions were similar when stratified by age <2 years (27% resistive and 73% elastic) and age >/= 2 years (32% resistive and 68% elastic).

Fig. 3 –

Comparison of contributions to mechanical power in adults with ARDS⁴ versus children with ARDS.

Discussion

In a large cohort of children with ARDS assessed under different ventilation flow modes, MP was marginally higher in decelerating (PCV or PRVC) compared to square (VCV) flow modes. We additionally characterized the contributions to MP in square flow ventilation and identified a resistive contribution of 30% that was independent of age when dichotomized as younger than two years old or greater than or equal to two years old.

To date, ventilator metrics have been inconsistently associated with outcome in pediatric ARDS, with residual confounding by severity of illness as higher settings are used in more severely ill children^15–19. This criticism also applies to many adult studies^10,20, as the association between metrics such as driving pressure and MP may themselves retain residual confounding due to severity of illness (e.g., worse elastance), and requires testing in a trial of different driving pressure or MP strategies to assess efficacy.

Decelerating flow modes are commonly used in pediatrics for myriad reasons including lower nominal PIPs and potentially improved oxygenation due to higher mean airway pressures⁸. The magnitude of the difference in MP between the modes was small, making the clinical significance of our finding unclear. In adults with ARDS, MP greater than 17 J/min has been associated with increased mortality and is a commonly-cited value³. However, pediatric definitions of MP require scaling tidal volumes to weight-based values given the wide range of age-related ventilator prescriptions in this patient population. This scaled MP value often makes it difficult to compare pediatric and adult data. A robust study pooling 4,549 adults with ARDS across six randomized clinical trials of protective mechanical ventilation identified a median weight-adjusted MP of 0.30 J·min−1 Kg−1 (IQR 0.13–0.64). Survivors had a median MP of 0.29 J·min−1 Kg−1 (IQR 0.12–0.59), and non-survivors had a median MP of 0.33 J·min−1 Kg−1 (IQR 0.14–0.72)⁴. The largest pediatric study evaluating MP reported on 546 children with ARDS and identified a median weight-adjusted MP of 0.53 (IQR 0.38, 0.74). Survivors had a median MP of 0.53 J·min−1 Kg−1 (IQR 0.37, 0.71), and non-survivors had a median MP of 0.61 J·min−1 Kg−1 (IQR 0.40–0.83)⁶. Another large pediatric study evaluating 306 children with ARDS reported an association between higher MP and fewer 28-day VFDs per 0.1 J·min−1 Kg−1⁵. Our findings that square flow ventilation is associated with 0.03 J·min−1 Kg−1 lower MP than decelerating flow ventilation in children with ARDS is likely not practice changing. Rather, it suggests that the use of decelerating flow, which is overwhelmingly (>95%) used in the pediatric population⁸, is likely not a priori exposing patients to increased risk of VILI compared to square flow ventilation. It also suggests that it is likely reasonable to use decelerating flow equations in most pediatric patients with confidence that MP approximates square flow-derived equations.

It was notable that MP was higher in children younger than two years old compared to older children, regardless of whether they were exposed to decelerating or square flow ventilation. Our data suggest that the higher respiratory rates required to maintain appropriate minute ventilation in children younger than two years old may be a primary contributor to this finding. When the respiratory rate component was removed from the MP equation (i.e. mechanical energy), there was more similarity between age cohort and mode of ventilation. The clinical significance of this finding is unclear. Our group previously reported the association between higher MP and worse clinical outcomes (mortality and 28-day VFDs) in a large cohort of children with pediatric ARDS⁶. In that study, given prior reports of age-based effects of MP and outcome, we explored whether there was effect modification by age less than 2 years and greater than or equal to 2 years. We did not find any effect modification by age on mortality or 28-day VFDs. In contrast, however, another recent large multicenter pediatric study⁵ concluded that higher MP is associated with fewer 28-day VFDs in children with ARDS, but in subgroup analysis by age, this association remained only in children < 2-years-old.

Patients who received continuous NMB infusions had higher MP in both square and decelerating flow than those patients who were not neuromuscularly blocked, and although not statistically significant in square flow, trended towards significance (p=0.0532). The higher MP seen in neuromuscularly blocked patients likely reflects a more severely ill cohort of patients who were placed under NMB, as patients under NMB had a higher mean PELOD-2 score (mean 7.70, SD 3.3, 95% CI 6.6–7.6) than those patients not under NMB (mean 6.43, SD 3.3, 95% CI 5.7–7.1) with a mean difference of 1.3 (95% CI 0.31–2.22) (p=0.01). The patients not under NMB were still breathing quietly (not above the ventilator), consistent with how larger adult studies have measured MP in unparalyzed ventilated subjects³.

Pediatric ARDS, unlike its adult counterpart, has a resistive component that is not easily quantified. The variable nature of decelerating flow modes makes it difficult to perform a bedside inspiratory hold maneuver to reliably calculate the resistive versus elastic contributions of the measured pressures. While our group has previously reported that PIP in decelerating flow overestimates plateau pressure in square flow by 1 to 2 cmH2O²¹, we have not directly evaluated the resistive contribution of MP in square flow in children with ARDS until this present study in a larger cohort. When we deconvoluted the relative contributions to MP in square flow VCV, we identified a resistive contribution of 30%, nearly three-fold higher than what has been reported in adults under identical VCV conditions (Figure 3)⁴.

The higher resistance in pediatric ARDS requires further exploration, as this may affect the utility of adult-based lung-protective strategies²². It is possible that the high resistance we describe in pediatrics may reflect a fundamental difference in pediatric ARDS respiratory physiology, compared to that of adult ARDS, potentially due to concurrent airway obstructive processes occurring alongside the parenchymal injury of ARDS. Indeed, as the nearly three-fold higher resistance in pediatrics compared to adults was maintained when we stratified our cohort by age < 2 and age >/= 2, it seems unlikely that we can attribute this higher resistance solely on diseases of younger childhood. It is also important to note that this observation does not argue that higher airway resistance seen in children is due to smaller endotracheal tubes than those used in adults. Poiseuille’s Law for laminar flow demonstrates that the smaller tube radius term in the denominator does not mathematically overcome the accompanying lower flow rates used in pediatrics. Despite nominally similar pressures as in adults when displayed on the ventilator, dissipation of these forces along a higher resistance airway may mitigate some of their potential for damage, although our data do not specifically support drawing this conclusion to change ventilator-prescription practice.

Given that most of the work investigating MP is derived from adult data, it is striking to evaluate our pediatric data against large numbers of adults with ARDS. In a large meta-analysis evaluating over 4,500 adult patients with ARDS, Costa et al.⁴ described summary data for common ventilator variables (Table 5). Notably, median PEEP (10 cmH2O) and median V_T (6 ml/kg) were identical in this adult study and in our pediatric cohort. Median plateau pressures were similar between populations (26 cmH2O in adults, 28 cmH2O in pediatrics). In contrast, median PIPs were higher in pediatrics (36 cmH2O) compared to adults (28 cmH2O). Median MP was also higher in pediatrics (46 ×100 J·min⁻¹ kg⁻¹) compared to adults (30 ×100 J·min⁻¹ kg⁻¹). Given the similarities between PEEP, tidal volumes, and respiratory rates between adults in the Costa et al. report and the children in our study, the higher MP we report in our pediatric cohort may primarily reflect the overall higher resistive PIP component we describe in children compared to adults, regardless of the child’s age.

Table 5 –

Comparison of demographic and ventilator variables in adults⁴ versus children with ARDS placed in square-flow volume-controlled ventilation. All data except for age are reported as median (Interquartile Range).

Ventilator Variables VCV/Square-Flow Ventilation	Adult ARDS4 (n=4,549)	Pediatric ARDS (n=185)
Age (years, mean (SD))	56 (24)	8.4 (6.7)
PEEP	10 (5, 20)	10 (10, 12)
Tidal Volume (mL/kg)	6 (4, 12)	6 (7, 8)
PIP	28 (18, 42)	36 (31, 41)
Plateau Pressure	26 (15, 41)	28 (24, 32)
Driving Pressure	14 (6, 29)	18 (14, 21)
Respiratory Rate	25 (11, 37)	24 (20, 30)
Age < 2 years	N/A	34 (29, 40)
Age >/= 2 years	N/A	20 (18, 25)
Mechanical Power	30 (13, 64)	46 (38, 74)

VCV (Volume-controlled ventilation), ARDS (Acute respiratory distress syndrome), PEEP (Positive end-expiratory pressure; cmH2O), PIP (Peak inspiratory pressure; cmH2O), Plateau Pressure (cmH2O), Driving Pressure (cmH2O), Mechanical Power (x100 J·Kg−1)

While our study demonstrates a 30% resistive contribution to MP in square flow ventilation, the remaining 70% is due to elastic properties including static (PEEP) and dynamic (Plateau minus PEEP). As PEEP contributes to both static and dynamic power, it should be a focus of future trials evaluating VILI in pediatric ARDS.

Our study has several limitations. It is a single-center cohort and generalizability to other centers may be limited. Additionally, our data only captured a single measurement at a single point in time, potentially under- or over-estimating our conclusions. Severity of illness and ARDS etiologies were similar to other reported cohorts²³, but management strategies such as use of ancillary therapies and non-conventional ventilation may affect our conclusions. The improved ventilation and perfusion that the lungs experience in prone positioning could potentially result in improved gas exchange, reduced resistance to air flow, and reduced MP. Like most adult studies, ours used only an inspiratory hold of 0.5 seconds to determine plateau pressure, but we acknowledge that longer inspiratory holds may permit more equilibration in heterogeneous lung injury. It should be noted, however, that a short inspiratory hold would overestimate the plateau pressure, and thus we may actually be reporting an underestimate of resistive contributions of MP. Some subjects (47%) did not undergo neuromuscular blockade, therefore measurements in these patients may not have been as reliable. However, our methodology is consistent with large adult studies, some of which also report an approximate 50% use of NMB and also assessed the absence of breathing effort in the non-NMB patients by comparing the set rate with the total respiratory rate in the ventilator³. Finally, we stress that the assumptions underlying the equations used in this study have only been assessed in adults, and it is possible they are not applicable to children. Indeed, pediatric ARDS may require its own MP formulations.

Our study has several strengths. The data was from a large, prospectively enrolled pediatric ARDS cohort from a large PICU with detailed data collection. Our sample size is one of the largest reported for pediatrics, making ours one of the few centers capable of providing a reasonably precise estimate of the differences in MP according to ventilator mode. Additionally, our measurements were made in the first 24 hours of ARDS, reflecting mostly acute pathophysiology, with potentially less confounding by co-interventions.

Our data support the need for further research in this area. The insights from our study could potentially provide justification for trials ultimately leading to modifications in ventilator protocols in pediatric intensive care units and in operating rooms. If our findings are validated in larger studies, it could lead to a shift towards individualized ventilator settings based on each patient’s unique respiratory mechanics, as determined by their MP. The potential for researching the impact of MP in the operating room is also significant. Intraoperative MP considerations could be particularly relevant during surgeries that involve dynamic changes in respiratory system mechanics and compliance, such as during certain thoracic (e.g. one-lung ventilation) and abdominal procedures. This could potentially reduce the incidence of intra- and postoperative pulmonary complications.

In conclusion, MP was lower in square versus decelerating flow ventilator modes, although the clinical significance of this is unclear. Approximately 30% of MP may go towards overcoming resistance, suggesting that some VILI may result from underappreciated properties such as flow rates.