Definition, Incidence, and Epidemiology

of Pediatric Acute Respiratory Distress

Syndrome: From the Second Pediatric Acute
Lung Injury Consensus Conference
Nadir Yehya, MD1,2
OBJECTIVES: In 2015, the Pediatric Acute Lung Injury Consensus Conference Lincoln Smith, MD3
(PALICC) provided the first pediatric-specific definitions for acute respiratory dis-
Neal J. Thomas, MD4
tress syndrome (pediatric acute respiratory distress syndrome [PARDS]). These
definitions have since been operationalized in cohort and interventional PARDS Katherine M. Steffen, MD5
studies. As substantial data have accrued since 2015, we have an opportunity to Jerry Zimmerman, MD3
assess the construct validity and utility of the initial PALICC definitions. Therefore, Jan Hau Lee, MD6
the Second PALICC (PALICC-2) brought together multiple PARDS experts and Simon J. Erickson, MD7
aimed to identify and summarize relevant evidence related to the definition and
Steven L. Shein, MD8
epidemiology of PARDS and create modifications to the definition of PARDS.
for the Second Pediatric
DATA SOURCES: MEDLINE (Ovid), Embase (Elsevier), and CINAHL Complete Acute Lung Injury Consensus
(EBSCOhost). Conference (PALICC-2) of the
STUDY SELECTION: We included studies of subjects with PARDS, or at risk Pediatric Acute Lung Injury and
Sepsis Investigators (PALISI)
for PARDS, excluding studies pertaining primarily to adults except as specified for Network
identifying age-specific cutoffs.
DATA EXTRACTION: Title/abstract review, full-text review, and data extraction
using a standardized data collection form.
DATA SYNTHESIS: The Grading of Recommendations Assessment,
Development, and Evaluation approach was used to identify and summarize ev-
idence and develop recommendations. A total of 97 studies were identified for
full-text extraction addressing distinct aspects of the PARDS definition, including
age, timing, imaging, oxygenation, modes of respiratory support, and specific
coexisting conditions. Data were assessed in a Patient/Intervention/Comparator/
Outcome format when possible, and formally summarized for effect size, risk, ben-
efit, feasibility of implementation, and equity. A total of 17 consensus-based defi-
nition statements were made that update the definition of PARDS, as well as the
related diagnoses of “Possible PARDS” and “At-Risk for PARDS.” These state-
ments are presented alongside a summary of the relevant epidemiology.
CONCLUSIONS: We present updated, data-informed consensus statements on
the definition for PARDS and the related diagnoses of “Possible PARDS” and
“At-Risk for PARDS.”
KEY WORDS: acute respiratory distress syndrome; epidemiologic factors;
pediatrics; ventilators mechanical

T
he 2015 Pediatric Acute Lung Injury Consensus Conference (PALICC)
provided the first definition of acute respiratory distress syndrome Copyright © 2023 by the Society of
(ARDS) specifically for pediatric patients (PARDS), formulated by Critical Care Medicine and the World
pediatric stakeholders (1). Prior to this, ARDS had been defined at the 1994 Federation of Pediatric Intensive and
American-European Consensus Conference (AECC) (2) and again in the 2012 Critical Care Societies
Berlin revision (3), primarily for adult patients by adult experts. The original DOI: 10.1097/PCC.0000000000003161

Pediatric Critical Care Medicine www.pccmjournal.org      S87

Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
 Yehya et al
PALICC definition of PARDS facilitated dedicated re-
search and subsequent studies that specifically tested
the utility of the newly developed definition. Since
then, large single- and multicenter observational stud-
ies, as well as randomized clinical trials, have used the
PALICC PARDS definition for enrollment.
As substantial data have accrued since 2015, the
Second PALICC (PALICC-2) was convened (4). In
this article, we address key question number 1 as out-
lined in the accompanying Methods article (4): “How
should PARDS be defined, and what are the variables
that best characterize the global burden of PARDS?”
Here, we detail the rationale for specific statements
underlying the 2022 PALICC-2 updated definition of
PARDS and the related diagnoses of “Possible PARDS”
and “At-Risk for PARDS.”
METHODS
The details of the literature search are outlined in the
PALICC-2 Methodology article in this supplement (4).
Due to the nature of the research questions, a scoping
review was conducted to identify relevant studies re-
lated to the topics that could inform the definition and
epidemiology of PARDS. These included studies re-
lated to occurrence rate and outcomes of patients with
PARDS, the association between age and PARDS ep-
idemiology or pathobiology, PARDS timing and trig-
gers, radiographic findings in PARDS, oxygenation
metrics, risk factors for development of PARDS, coex-
istence of PARDS and cardiac disease, and PARDS in
patients with congenital heart disease and chronic pul-
monary failure. Adult data were excluded except when
considering the association between age (up to 40 yr
old) and ARDS diagnosis and epidemiology. The com-
plete search strategies can be found in Supplemental
Tables 1–5 (http://links.lww.com/PCC/C296). Details
of title/abstract review, full-text review, data extrac-
tion, and generation of clinical practice recommenda-
tions, research statements, and policy statements are
outlined in the PALICC-2 Methodology article (4). In
general, revisions of the original PALICC definition
were only considered when supported by new data.
RESULTS
Results of the literature searches are detailed in the
Supplemental Digital Content (http://links.lww.com/
PCC/C296) with the Preferred Reporting Items for
Systematic Reviews and Meta-Analyses flowcharts for
S88      www.pccmjournal.org February 2023 • Volume 24 • Number 2 (Suppl 1)
Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
each of the corresponding subsections detailed below
(Supplemental Figs. 1–6, http://links.lww.com/PCC/
C296). As detailed below (and in Tables  1 and 2),
children on invasive mechanical ventilation (IMV) or
full-face noninvasive ventilation (NIV) can be diag-
nosed with PARDS when hypoxemia is sufficient or be
termed at-risk for PARDS when it is not. Children on
nasal support modes can be diagnosed with possible
PARDS when hypoxemia is sufficient or be termed at-
risk for PARDS when it is not.
Age
There may be age-related differences in the epidemi-
ology of PARDS, and there are existing definitions for
ARDS in neonates (5) and adults (Berlin) (3) that dif-
fer from PALICC criteria (e.g., requiring bilateral infil-
trates). However, there are insufficient data to justify
precise population-level, age-based cutoffs to move
between the published definitions of neonatal ARDS
(NARDS), PARDS, and ARDS among patients cared
for pediatric intensivists who may range in age from
neonates to young adults.
Definition Statement 1.1. All patients less
than 18 years of age without active perinatal
lung disease should be diagnosed with PARDS
using PALICC-2 criteria.
Remarks: Practitioners can use either the
PALICC-2 or neonatal definition (Montreux
NARDS) for neonates and can use either the
PALICC-2 or adult definition (Berlin ARDS)
for young adults (Ungraded definition state-
ment, 94% agreement).
Justification. We found no studies establishing age-
dependent differences between neonates and children
on the epidemiology of ARDS (6, 7). Some data sup-
port changes in epidemiology and outcomes of ARDS
along the spectrum of ages typically seen in the PICU
(8–10). In an observational study of children with
acute lung injury in China, ARDS (AECC criteria) was
associated with older age (4.8 yr [interquartile range
0.5–8 yr] vs 0.6 yr [0.3–1 yr]; p = 0.01) (8). Older age
was also associated with PARDS among PICU trauma
patients (9, 10). However, three other studies of ARDS
in children did not suggest age-dependent differences
in severity or outcomes (11–13). Similarly, we found
no data supporting a specific age cutoff among younger
adults, with an insufficient number of subjects between
 PALICC-2 Supplement

TABLE 1.
Diagnosis of Pediatric Acute Respiratory Distress Syndrome (Definition Statement 1.7.1)
Variable Definition

Age (DS 1.1) Exclude patients with perinatal lung disease

Timing (DS 1.2)
Origin of pulmonary
edema (DS 1.3)
Chest imaging (DS 1.3)
Oxygenationb (DS 1.4.3)
Special populationsd
Cyanotic heart disease
(DS 1.6.1, 1.6.2)
Chronic lung disease
(DS 1.6.3, 1.6.4)
Within 7 d of known clinical insult
Not fully explained by cardiac failure or fluid overload
New opacities (unilateral or bilateral) consistent with acute pulmonary parenchymal disease and
which are not due primarily to atelectasis or pleural effusiona
IMV: OI ≥ 4 or OSI ≥ 5
NIV: Pao2/Fio2 ≤ 300 mm Hg or Spo2/Fio2 ≤ 250
Stratification of PARDS Severity: Apply ≥ 4 hr after initial diagnosis of PARDS (DS 1.4.4)
IMV-PARDS: Mild/moderate PARDS: OI < 16 Severe PARDS: OI ≥ 16 or
(DS 1.4.1) or OSI < 12 OSI ≥ 12
NIV-PARDS:c Mild/moderate NIV-PARDS: Pao2/Fio2 > Severe NIV-PARDS: Pao2/Fio2
(DS 1.4.2) 100 or Spo2/Fio2 > 150 ≤ 100 or Spo2/Fio2 ≤ 150
Above criteria, with acute deterioration in oxygenation not explained by cardiac disease
Above criteria, with acute deterioration in oxygenation from baseline

DS = definition statement, IMV = invasive mechanical ventilation, NIV = noninvasive ventilation, OI = oxygenation index, OSI = oxygen
saturation index, PARDS = pediatric acute respiratory distress syndrome, Spo2 = oxygen saturation.
a
Children in resource-limited environments where imaging is not available who otherwise meet PARDS criteria are considered to have
possible PARDS.
b
Oxygenation should be measured at steady state and not during transient desaturation episodes. When Spo2 is used, ensure that Spo2 ≤
97% (DS 1.4.5).
OI = mean airway pressure (MAP) (cm H2O) × Fio2/Pao2 (mm Hg).
OSI = MAP (cm H2O) × Fio2/Spo2.
c
Diagnosis of PARDS on NIV (NIV-PARDS) requires full facemask interface with continuous positive airway pressure (or expiratory
positive airway pressure) ≥ 5 cm H2O.
d
Stratification of PARDS severity does not apply to these populations.
PARDS should not be diagnosed in children with respiratory failure solely from airway obstruction (e.g., critical asthma, virus-induced
bronchospasm).

18 and 40 years old included in adult studies. In adults bilateral opacities). On balance, the consensus was that
with hypoxemic respiratory failure, patients with the PALICC-2 consensus statements are intended for
greater than or equal to 1 ARDS risk factor were older children less than 18 years without acute perinatal condi-
(62 ± 19 vs 49 ± 22 yr) and had a higher in-hospital tions (e.g., meconium aspiration, surfactant deficiency)
mortality (33.9% vs 17.3%) as compared with patients and are recommended for use in that population but that
with no ARDS risk factors (14), suggesting that even in practitioners may use their preferred definitions for neo-
adults, worse prognosis with age may be due primarily nates (PALICC-2 or Montreux) or teenagers and young
to accumulation of comorbidities. adults (PALICC-2 or Berlin). This may negatively impact
The benefit of not having specific age cutoffs for definition implementation, in particular, inhibiting au-
PARDS is that practitioners who care for neonates, chil- tomated systems that require clear, objective definitions.
dren, and/or adults can use the definition of ARDS for
which they are the most comfortable for infants, teen-
Timing and Triggers
agers, and young adults. Potential challenges include
misclassification of patients especially related to key Since the 2012 Berlin definition, “acute” onset of hypox-
differences between the definitions (e.g., unilateral vs emia has been operationalized as occurrence within 7
Pediatric Critical Care Medicine www.pccmjournal.org      S89
Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
 Yehya et al

TABLE 2.
Diagnosis of Possible Pediatric Acute Respiratory Distress Syndrome and At-Risk for
Pediatric Acute Respiratory Distress Syndrome (Definition Statement 1.5.3, 1.7.2)
Variable Definition

Age Exclude patients with perinatal lung disease

Timing Within 7 d of known clinical insult
Origin of pulmonary edema Not fully explained by cardiac failure or fluid overload
Chest imaging New opacities (unilateral or bilateral) consistent with acute pulmonary parenchymal
disease and which are not due primarily to atelectasis or effusiona
Oxygenationb threshold to diagnose possible PARDS for children on nasal respiratory supportc (definition statement 1.5.1)
Nasal continuous positive airway pressure/bilevel positive airway pressure or high-flow
nasal cannula (≥ 1.5 L/kg/min or ≥ 30 LPM): Pao2/Fio2 ≤ 300 or Spo2/Fio2 ≤ 250
Oxygenationb threshold to diagnose at-risk for PARDS
Any interface: Oxygen supplementationd to maintain Spo2 ≥ 88 but not meeting defini-
tion for PARDS or possible PARDS (see above)
Special populations
 Cyanotic heart disease Above criteria, with acute deterioration in oxygenation not explained by cardiac disease
 Chronic lung disease Above criteria, with acute deterioration in oxygenation from baseline

PARDS = pediatric acute respiratory distress syndrome, Spo2 = oxygen saturation.

a
Children in resource-limited environments where imaging is not available who otherwise meet possible PARDS criteria are considered
to have possible PARDS (definition statement 1.5.2).
b
Oxygenation should be measured at steady state and not during transient desaturation episodes. When Spo2 is used, ensure that Spo2 ≤ 97%.
c
Children on nasal noninvasive ventilation (NIV) or high-flow nasal cannula are not eligible for PARDS but are considered to have
possible PARDS when this oxygenation threshold is met.
d
Oxygen supplementation is defined as Fio2 > 21% on invasive mechanical ventilation; or Fio2 > 21% on NIV; or “oxygen flow” from a
mask or cannula that exceeds these age-specific thresholds: ≥ 2 L/min (age < 1 yr), ≥ 4 L/min (age 1–5 yr), ≥ 6 L/min (age 6–10 yr), or
≥ 8 L/min (age > 10 yr). For children on a mask or cannula, oxygen flow calculated as Fio2 × flow rate (L/min) (e.g., 6 L/min flow at 0.35
Fio2 = 2.1 L/min).
Possible PARDS and at-risk for PARDS should not be diagnosed in children with respiratory failure solely from airway obstruction (e.g.,
critical asthma, virus-induced bronchospasm).

days of an inciting trigger (5). This cutoff was adopted
by PALICC in 2015 for PARDS (1). We identified no
studies of PARDS in which cases were attempted to
be identified beyond 7 days after an identified trigger,
suggesting widespread acceptance of this threshold.
Definition Statement 1.2. Symptoms of hy-
poxemia and radiographic changes must occur
within 7 days of a known insult to qualify for
PARDS (Ungraded definition statement, 96%
agreement).
Justification. There is widespread acceptance of 7
days as the cutoff between PARDS trigger and meet-
ing the remainder of the criteria (hypoxemia and im-
aging), such that nearly all cohort studies have adopted
this threshold. In all large cohort studies reviewed,
PARDS was either present at PICU admission or was
diagnosed within 24 hours for most subjects (12, 15,
16). The median time between intubation and PARDS
diagnosis was less than 24 hours in all reviewed stud-
ies. The benefits of keeping the window up to 7 days
are consistency with existing literature and sensitivity,
as it allows subjects with risk factors to plausibly de-
velop the clinical findings of PARDS in a reasonable
timeframe. Shortening (e.g., within 3 d) would plau-
sibly capture the majority of subjects but risks missing
some patients who were still developing the specific
combination of clinical and radiographic criteria for
PARDS. On balance, the increased sensitivity, congru-
ence with existing literature and adult Berlin ARDS,
and ease of operationalizing this was felt to outweigh
any benefits to shortening the observation window.

S90      www.pccmjournal.org February 2023 • Volume 24 • Number 2 (Suppl 1)

Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited

Imaging Findings
Bilateral pulmonary opacities are required to diagnose
ARDS in adults and neonates, while the 2015 PALICC
required solely the presence of a “new” opacity, irre-
spective of whether this was isolated to a single lobe,
quadrant, or lung. The rationale for this departure
related to the sensitivity and timing of chest radiog-
raphy to detect a diffuse inflammatory process such
as PARDS, the inter-rater reliability of chest radiog-
raphy, and the lack of prognostic relevance of bilateral
opacities on PARDS outcomes after controlling for
hypoxemia.
Definition Statement 1.3. Chest imaging find-
ings of new opacity (or opacities) consistent
with acute pulmonary parenchymal disease
not explained by atelectasis or effusion are nec-
essary to diagnose PARDS. (Ungraded defini-
tion statement, 90% agreement).
Justification. Since 2015, multiple studies have di-
rectly or indirectly examined chest imaging in PARDS.
In the multinational Pediatric Acute Respiratory
Distress Syndrome Incidence and Epidemiology
(PARDIE) study (12), inter-rater reliability for deter-
mination of bilateral opacifications was very poor, with
the kappa comparing pediatric intensivists and pedi-
atric radiologists only 0.31 (17). In addition, bilateral
infiltrates were eventually scored in 87% of all subjects
with PARDS, suggesting that bilateral infiltrates may
lag as a marker. Finally, bilateral infiltrates at PARDS
onset were not independently associated with mortality
in PARDIE (12). A single study using a binary bilateral
versus not bilateral definition suggested lower mor-
tality with bilateral infiltrates (18), further highlight-
ing the inconsistent prognostic relevance of unilateral
versus bilateral opacities in the definition. Hence, there
was no new evidence contrary to the rationale used in
PALICC for simplification of the radiographic criteria.
Benefits of radiographic criteria that do not require
bilateral opacifications include more reproducible
and earlier recognition of PARDS, improving sensi-
tivity. Potential challenges of not requiring bilateral
opacifications include potential for lower specificity
in the diagnosis of PARDS. However, on balance, the
group valued earlier and more complete recognition
of patients with PARDS. Furthermore, while most
data are generated from chest radiographs, the panel
Pediatric Critical Care Medicine www.pccmjournal.org      S91
Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
PALICC-2 Supplement
considered it important to allow for the inclusion of
other methods of lung imaging to diagnose PARDS
(such as CT or lung ultrasound). The use of lung ul-
trasound, in particular, may be important in lower re-
source settings, where chest radiography may not be
readily available.
Severity of Hypoxemia for Disease Stratification
In 2015, PALICC established severity categories
according to degree of hypoxemia and type of support,
including NIV PARDS, which was not further strati-
fied by severity, and invasive (intubated) PARDS strat-
ified as mild, moderate, or severe according to degree
of hypoxemia at PARDS onset. Both Pao2 and oxygen
saturation (Spo2)-based metrics were used: Pao2/Fio2
and Spo2/Fio2 were used in NIV PARDS, and oxygen-
ation index (OI)/oxygen saturation index (OSI) used
to stratify severity in intubated patients with PARDS.
Since 2015, large cohort studies have assessed the
utility of OI/OSI for intubated children, identified
prognostic significance to degree of hypoxemia during
NIV PARDS, and documented improved risk stratifi-
cation after a period of stabilization.
Definition Statement 1.4.1. OI or OSI, in
preference to Pao2/Fio2 or Spo2/Fio2, should
be the primary metric of lung disease severity
to define PARDS for all patients treated with
IMV, with Pao2 used preferentially when avail-
able. (Ungraded definition statement, 90%
agreement).
Definition Statement 1.4.2. Pao2/Fio2 or Spo2/
Fio2 should be used to diagnose PARDS and
possible PARDS for patients receiving NIV or
high flow nasal cannula. (Ungraded definition
statement, 88% agreement).
Definition Statement 1.4.3. Patients on a
full-face interface of NIV [continuous posi-
tive airway pressure (CPAP) or bilevel posi-
tive airway pressure (BiPAP)] with CPAP ≥
5 cm H2O or IMV should be considered as
having PARDS if they meet timing, oxygen-
ation, etiology/risk factor, and imaging cri-
teria (Ungraded definition statement, 90%
agreement).
Definition Statement 1.4.4. Subjects with
PARDS should be stratified into severity
 Yehya et al
categories after a period of at least 4 hours.
(Ungraded definition statement, 85%
agreement).
Definition Statement 1.4.5. When apply-
ing Spo2 criteria to diagnose PARDS, oxygen
should be titrated to achieve an Spo2 between
88% and 97%. (Ungraded definition statement,
96% agreement).
Justification. Incorporating mean airway pressure
into the hypoxemia metric (i.e., OI and OSI) consist-
ently, albeit marginally, improved risk stratification in
PARDS compared with Pao2/Fio2 or Spo2/Fio2 (12,
15, 19). In a single-center study assessing hypoxemia
metrics at PARDS onset and again at 24 hours after,
Pao2/Fio2 had area under the receiver operating char-
acteristic (AUROC) curve of 0.56 to 0.66 for mortality,
relative to AUROCs of 0.62 to 0.71 with OI/OSI at com-
parable time points, although this was partly due to
greater inclusion within PALICC by use of Spo2-based
OSI (15). In PARDIE, PALICC stratification using OI
and OSI had better mortality discrimination at PARDS
onset (p = 0.09) and at 6, 12, 18, and 24 hours (all p <
0.05), relative to Berlin stratification using Pao2/Fio2
and Spo2/Fio2 (12). Finally, in a separate cohort, Pao2/
Fio2 and OI discriminated mortality similarly, but OI
better correlated with nonpulmonary organ failures
(19). The risks of keeping OI/OSI for intubated sub-
jects, rather than Pao2/Fio2 and Spo2/Fio2 for both
NIV and IMV PARDS, is increased complexity of hav-
ing two separate metrics and the potentially erroneous
assumptions surrounding the accuracy of measuring
mean airway pressure in different modes of ventila-
tion. Overall, the balance was felt to favor keeping sep-
arate hypoxemia metrics for NIV and IMV PARDS.
The benefit of including Spo2-based criteria is
increased sensitivity, with substantially more subjects
meeting PARDS criteria and meeting criteria earlier
(12, 15), provided that Spo2 is lower than the plateau
portion of the oxyhemoglobin dissociation curve
(80–97%). This increased sensitivity was considered
sufficient to justify keeping Spo2-based oxygenation
metrics, particularly given variable utilization of arte-
rial blood gasses. The risk of Spo2-based metrics, in-
cluding potential inaccuracies according to race and
perfusion, were considered, but the balance was con-
sidered to favor continued use of OSI and Spo2/Fio2.
We found improved mortality discrimination sev-
eral hours after the initial qualifying hypoxemia metric
S92      www.pccmjournal.org February 2023 • Volume 24 • Number 2 (Suppl 1)
Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
Figure 1. Area under the receiver operating characteristic
curves plot for the utility of Second Pediatric Acute Lung Injury
Consensus Conference categories (mild, moderate, severe) to
discriminate PICU mortality at pediatric acute respiratory distress
syndrome onset and at 6, 12, 18, and 24 hr after onset in a cohort
composed of previously published single- and multicenter cohorts.
was recorded (12, 16, 20). PARDIE and a single-cen-
ter study from a quaternary center demonstrated that
measuring hypoxemia starting at approximately 6
hours after PARDS onset better discriminated PICU
mortality, both individually (12, 16, 21) and combined
(Fig. 1). As these studies allowed some hypoxemia
metrics recorded within 2 hours of their initial diag-
nostic timestamp, and because 6 hours can be difficult
to operationalize for trial purposes, we suggest strati-
fying PARDS severity after at least 4 hours of standard
management after the initial qualifying hypoxemia
metric. The benefits of delayed severity stratification
include improvement in mortality discrimination and
minimizing exposure to unnecessary therapies in chil-
dren who improve quickly following PARDS diagnosis
(e.g., with lung recruitment following endotracheal
intubation). Risks include increased workload and
the possibility of delays in aggressive care for children
who would benefit from them. Some ongoing obser-
vational studies (NCT04113434) and clinical trials
(NCT03896763) have already instituted 4-hour peri-
ods of stabilization prior to eligibility, demonstrating
feasibility of this approach for at least research pur-
poses. Overall, the balance for accurately identifying
higher risk subjects, particularly for appropriate use of
more aggressive therapies, outweighed the potential
drawbacks.

Figure 2. Mortality rates according to 2015 Pediatric Acute Lung
Injury Consensus Conference severity cutoffs from three cohorts
(Pediatric Acute Respiratory Distress Syndrome Incidence and
Epidemiology, cohort from Children’s Hospital of Philadelphia,
cohort from Children’s Hospital of Los Angeles) using values
between 6 and 12 hr (total n = 1149). Note the large increase in
mortality with severe pediatric acute respiratory distress syndrome.
OI = oxygenation index.
Given the lack of prognostic distinction between
the previously categorized mild and moderate PARDS
(Fig. 2) for both NIV and intubated subjects, we have
simplified to binary severity categorization: mild/
moderate and severe. For both IMV (using PALICC
OI/OSI cutoffs) and NIV (using Berlin Pao2/Fio2 and
Spo2/Fio2 cutoffs), there was a large increase in mor-
tality in the severe category, with little discrimination
between mild and moderate, particularly in PARDIE
(12). Similar patterns were seen in large single-center
studies (15, 16). The benefits of this approach are sim-
plicity and an accurate categorization according to
mortality risk. The disadvantage is a loss of discrimi-
nation for nonmortality endpoints, such as ventilator
duration in survivors or ventilator-free days, which
did have appropriate calibration in a three-level (mild,
moderate, severe) stratification system. A binary strat-
ification system is also divergent from NARDS and
Berlin ARDS definitions, both of which use a three-
level stratification system. On balance, the simplicity
Pediatric Critical Care Medicine www.pccmjournal.org      S93
Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
PALICC-2 Supplement
and sufficient accuracy of a binary severity classifica-
tion system separately for IMV and NIV PARDS was
favored.
Possible PARDS and At-Risk for PARDS
In 2015, the novel category of “at-risk for PARDS” was
created by PALICC as a cohort of children who did not
meet PARDS criteria but warranted research aimed
at understanding and preventing disease progression.
Children on full-face NIV or on IMV with a PARDS
trigger and chest imaging consistent with PARDS were
diagnosed with at-risk for PARDS if they exhibited
hypoxemia that was too mild to meet PARDS crite-
ria (e.g., OI < 4). Full-face NIV implies any interface
that covers the face and mouth, including oro-nasal
and Scuba-style masks. The at-risk for PARDS defini-
tion included children on nasal modes of respiratory
support (high-flow nasal cannula [HFNC] or nasal
CPAP), as they were deemed ineligible for PARDS due
to unquantifiable degree of entrainment of room air,
thus making calculations of Pao2/Fio2 or Spo2/Fio2
unreliable. Children in this latter group could plau-
sibly have PARDS-level hypoxemia but only could be
diagnosed with PARDS if a clinician changed their res-
piratory support interface. We have now aimed to sep-
arate these patients into two distinct groups: those that
may have PARDS but cannot be diagnosed with it due
to interface selection or imaging availability (Possible
PARDS), and those that do not have PARDS based on
mild hypoxemia but may develop it given the presence
of all other PARDS criteria (at-risk for PARDS).
Definition Statement 1.5.1. Patients on a
nasal interface of NIV (CPAP or BiPAP) or on
HFNC ≥ 1.5 L/kg/min or ≥ 30 L/min should be
considered as having possible PARDS if they
meet timing, oxygenation, etiology/risk fac-
tor, and imaging criteria (Ungraded definition
statement, 87% agreement).
Definition Statement 1.5.2. Patients who
do not have imaging due to resource limita-
tions should be considered as having possible
PARDS if they otherwise meet timing, oxygen-
ation, and risk factor criteria. (Ungraded defi-
nition statement, 90% agreement).
Definition Statement 1.5.3. Defining a
group of patients at risk for PARDS is neces-
sary to determine the epidemiology of disease
 Yehya et al
progression and potential avenues for disease
prevention. (Ungraded definition statement,
96% agreement).
Justification. The original at-risk for PARDS crite-
ria were validated in the PARDIE study and a single-
center study of children with viral lower respiratory
tract infections, both of which found an increased rate
of subsequent PARDS among children meeting at-risk
for PARDS criteria (22, 23). It is not known how many
at-risk for PARDS patients would have been diagnosed
with PARDS had a clinician changed their interface (or
level of respiratory support). To test the potential im-
pact of these new definitions of possible PARDS, we
performed secondary analysis of the PARDIE dataset
to identify support levels that discriminated children
with less favorable outcomes as a surrogate for un-
diagnosed PARDS. Of the 222 patients who met the
PALICC 2015 definitions for at risk for PARDS, 50
(22.5%) would now be reclassified as meeting the pro-
posed possible PARDS definition, all but one of which
(n = 49) were on HFNC (Supplemental Table 6, http://
links.lww.com/PCC/C296). Compared with children
not meeting possible PARDS criteria, categorization as
possible PARDS was associated with a two-fold higher
percentage of subsequent PARDS (37% vs 17%; p =
0.007), and the 31 possible PARDS patients who were
not subsequently diagnosed with PARDS had a longer
PICU length of stay than the children who never met
possible PARDS or PARDS criteria (5.3 d [3.2–8.3 d]
vs 3.0 d [1.7–5.1 d]; p < 0.001) (Supplemental Table
6, http://links.lww.com/PCC/C296). The Possible
PARDS definition is further supported by recent data
showing most children with bronchiolitis (71%) on
HFNC had a peak inspiratory flow of less than 1.5–2 L/
min/kg. This suggests that significant entrainment of
room air is unlikely for children on HFNC supported
with HFNC at flow rates greater than or equal to 1.5 L/
kg/min, indicating that prescribed Fio2 is likely sim-
ilar to delivered Fio2 and is reasonable for calculating
Spo2/Fio2 ratio (24). Additionally, experts in ARDS
suggested making adults on HFNC of greater than
or equal to 30 L/minute eligible to be diagnosed with
ARDS (25), and pediatric experts recently defined res-
piratory failure using our same HFNC thresholds of
greater than or equal to 1.5 L/kg/min or greater than
or equal to 30 L/min (26).
The benefits of labeling children with significant hy-
poxemia as “Possible PARDS” include calling attention
S94      www.pccmjournal.org February 2023 • Volume 24 • Number 2 (Suppl 1)
Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
to their risk of poor outcomes and possible need for
closer monitoring (PICU instead of general ward),
avoidance of additional injurious stimuli, and provi-
sion of possibly beneficial treatments. Alternatively,
the possible PARDS definition could cause harm by
increasing unnecessary PICU utilization or exposing
moderately ill children to unproven PARDS-related
therapies. There are likely differences between nasal
possible PARDS patients and those missing chest im-
aging, but the consensus was to use one term to describe
all patients who possibly have PARDS but are unable to
be properly diagnosed; clinicians should consider this
when deciding how to treat possible PARDS patients.
Subjects diagnosed with possible PARDS due to using
a nasal interface could be escalated to full-face NIV to
assess whether they are most appropriately stratified as
having PARDS or being at-risk for PARDS, and fur-
ther research will hopefully enable better stratification
in the next iteration of the guidelines. On balance, the
group favored creation of this new definition to iden-
tify children who may benefit from closer monitoring
and require further research aimed at preventing pro-
gression to PARDS, especially given that implemen-
tation of the new definition of possible PARDS was
felt to be reasonably feasible. Similarly, the availability
and indications for chest imaging vary substantially
according to geography and economics, neither of
which impact the underlying pathophysiology or de-
gree of hypoxemia. Including children lacking imaging
due to resource limitations in this new diagnosis is an
advantage that promotes equity in the applicability of
our definition.
Diagnosing PARDS in Patients With Chronic
Cardiorespiratory Disease
The 2015 definition permitted diagnosis of PARDS in
children with coexisting left ventricular (LV) dysfunc-
tion, cyanotic congenital heart disease, and chronic
lung disease requiring variable levels of support. In rec-
ognizing the difficulty of risk stratifying these subjects
due to unreliable metrics of hypoxemia (cyanotic mix-
ing lesions) or because of variable baselines (chronic
invasive ventilation), these subjects are not risk strati-
fied using hypoxemia metrics. Our scoping review did
not reveal new literature supporting changes in these
recommendations, all of which were reapproved by the
PALICC-2 members with greater than or equal to 90%
agreement.

Definition Statement 1.6.1. Patients with cy-
anotic congenital heart disease are considered
to have PARDS if they fulfill standard PARDS
criteria and have an acute deterioration in
oxygenation (relative to baseline) not fully
explained by the underlying cardiac disease.
(Ungraded definition statement, 98%
agreement).
Definition Statement 1.6.2. Patients with LV
dysfunction who fulfill standard PARDS cri-
teria are considered to have PARDS if acute
hypoxemia and new chest imaging changes
cannot be explained solely by LV heart failure
or fluid overload. (Ungraded definition state-
ment, 92% agreement).
Remarks: Cardiac ultrasonography and/or left
atrial pressure may be useful in identifying hy-
drostatic pulmonary edema
Definition Statement 1.6.3. Patients with
preexisting chronic lung disease treated with
supplemental oxygen, NIV, or IMV via trache-
ostomy are considered to have PARDS if they
demonstrate acute changes that fulfill standard
PARDS criteria and exhibit an acute deteriora-
tion in oxygenation from baseline that meets
oxygenation criteria for PARDS.
(Ungraded definition statement, 96%
agreement).
Definition Statement 1.6.4. Patients with
chronic lung disease who receive mechanical
ventilation at baseline or have cyanotic con-
genital heart disease with acute onset of illness
that fulfill standard PARDS criteria should
not be stratified by OI or OSI risk categories.
Future studies are necessary to determine
PARDS risk stratification of such patients
with acute-on-chronic hypoxemic respiratory
failure. (Ungraded definition statement, 90%
agreement).
Justification. Hydrostatic pulmonary edema (LV
dysfunction, fluid overload, pulmonary lymphatic
failure) and alveolar-pulmonary capillary membrane
injury can occur individually and concomitantly.
Similarly, acute and chronic lung disease frequently
coexist. On balance, classifying children with cyanotic
congenital heart disease, LV dysfunction, right-to-left
intracardiac shunting, total body fluid overload, and
chronic lung disease with acute illness that otherwise
Pediatric Critical Care Medicine www.pccmjournal.org      S95
Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
PALICC-2 Supplement
satisfy PARDS criteria as having PARDS was felt to be
beneficial, because the underlying pathophysiology was
similar and such subjects likely benefit from PARDS-
directed therapies. Finally, stratifying chronically in-
vasively ventilated subjects and children with cyanotic
lesions based on OI or OSI should be avoided. Variable
baselines among chronically ventilated subjects make
using standard OI/OSI PARDS cutoffs potentially mis-
leading. Subjects with chronic lung disease normally
supported on NIV, subjects with noncyanotic heart
lesions (including palliated and surgically corrected
lesions), and those with LV dysfunction can be strat-
ified using OI and OSI if intubated.
Definition of PARDS, Possible PARDS, and
At-Risk for PARDS
The above statements were formalized into the tables
defining PARDS (Table  1) and Possible PARDS and
at-risk for PARDS (Table 2). We also provide a flow-
chart for stratifying subjects according to mode of sup-
port and degree of hypoxemia (Fig. 3). PARDS can be
diagnosed only in children on IMV or full-face NIV.
Possible PARDS can be diagnosed for children on
nasal modes (and IMV or full-face NIV in resource-
limited environments where chest imaging is una-
vailable). Children on any mode who meet PARDS
criteria except have insufficient hypoxemia are termed
at-risk for PARDS. Compared with the 2015 tables, we
clarified that atelectasis and pleural effusion are not
qualifying opacities. The threshold for NIV PARDS by
Spo2/Fio2 was rounded to 250 for ease of implemen-
tation and deemed acceptable given lack of evidence
supporting either cutoff as superior. New footnotes
specifically excluding acute respiratory failure solely
from obstructive processes (e.g., critical asthma, virus-
induced bronchospasm) were added. LV dysfunction
was removed from special populations as it was redun-
dant with the origin of edema.
Definition Statement 1.7.1. PARDS shall be
defined using Table  1 (Ungraded definition
statement, 84% agreement).
Definition Statement 1.7.2. Possible and
at risk for PARDS shall be defined using
Table  2 (Ungraded definition statement, 94%
agreement).
Remarks: Clinicians should consider treating
patients with possible PARDS as if they have
 Yehya et al

Figure 3. Flowchart for diagnosing pediatric acute respiratory distress syndrome (PARDS), possible PARDS, and at-risk for PARDS
based on type of support and degree of hypoxemia. HFNC = high-flow nasal cannula, IMV = invasive mechanical ventilation,
LPM = L/minute, NIV = noninvasive ventilation, OI = oxygenation index, OSI = oxygen saturation index, P/F = ratio of Pao2/Fio2, S/F =
ratio of oxygen saturation/Fio2, Spo2 = oxygen saturation.

PARDS and apply other recommendations
after considering the specific risks and benefits
for that specific patient.
Justification. Most elements of the definitions are
discussed in the Justifications above. It is notable that
Definition Statement 1.7.1 met our a priori criterion
for acceptance but had the lowest agreement (84%) of
all statements in this section. This is likely because a
panelist disagreeing with any of the other statements
(e.g., age, chest imaging) would likely disagree with
that element of Table 1, in essence making all disagree-
ments funnel to this Statement. In particular, changes
in severity stratification to two categories (from three
[IMV] and one [NIV] in PALICC-1) and to a delayed
strategy (i.e., at ≥ 4 hr after PARDS diagnosis) were
most debated among the final statements, although
they were retained based on the justifications described
above.
CONCLUSIONS
In 2015, we provided the first pediatric-specific defini-
tion of PARDS for children. We now provide a revised
definition of PARDS based on current data and ex-
pert consensus. The major changes to the definition of
PARDS revolve around the timing and severity strati-
fication of hypoxemia, with the institution of a 4-hour
waiting period prior to risk stratification and a simpli-
fied stratification of nonsevere or severe for both NIV
PARDS and invasively ventilated PARDS. Furthermore,
we have introduced the concept of Possible PARDS
to capture subjects who likely possess underlying
pathophysiology consistent with PARDS, including

S96      www.pccmjournal.org February 2023 • Volume 24 • Number 2 (Suppl 1)

Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited

children missing imaging in resource-limited environ-
ments. While controversies remain, particularly in the
domains of age, PARDS trajectory, and noninvasive
interfaces, we propose using this definition for clinical
practice and for future studies of PARDS, with a re-
search focus on remaining areas of controversy to con-
tinue to improve and refine the definition.
1
Division of Pediatric Critical Care, Department of
Anesthesiology and Critical Care, Children’s Hospital of
Philadelphia and University of Pennsylvania, Philadelphia, PA.
2 Leonard Davis Institute of Health Economics, University of
Pennsylvania, Philadelphia, PA.
3 Seattle Children’s Hospital and Harborview Medical Center,
Division of Pediatric Critical Care Medicine, Department of
Pediatrics, University of Washington, Seattle, WA.
4 Division of Pediatric Critical Care Medicine, Department of
Pediatrics and Public Health Science, Penn State Hershey
Children’s Hospital, Hershey, PA.
5 Division of Pediatric Critical Care, Department of Pediatrics,
Stanford University, Palo Alto, CA.
6 Children’s Intensive Care Unit, KK Women’s and Children’s
Hospital, Singapore.
7 Department of Paediatric Critical Care, Perth Children’s
Hospital and University of Western Australia, Perth, WA,
Australia.
8 Division of Pediatric Critical Care Medicine, Department of
Pediatrics, Rainbow Babies and Children’s Hospital and
Case Western Reserve University, Cleveland, OH.
Supplemental digital content is available for this article. Direct
URL citations appear in the printed text and are provided in the
HTML and PDF versions of this article on the journal’s website
(http://journals.lww.com/pccmjournal).
All authors were involved in the conception, design, and drafting
of the article.
Dr. Yehya’s institution received funding from the National Heart,
Lung, and Blood Institute; he received support for article re-
search from the National Institutes of Health. Dr. Lee’ s institution
received funding from the National Research Medical Council,
Singapore. Dr. Shein received funding from Hill Ward Henderson
Law Firm. The remaining authors have disclosed that they do not
have any potential conflicts of interest.
For information regarding this article, E-mail: yehyan@email.
chop.edu
The Second Pediatric Acute Lung Injury Consensus Conference
(PALICC-2) group members are listed in Appendix 1 (http://
links.lww.com/PCC/C296).
REFERENCES
1. Khemani RG, Smith LS, Zimmerman JJ, et al; Pediatric Acute
Lung Injury Consensus Conference Group: Pediatric acute
respiratory distress syndrome: Definition, incidence, and
epidemiology: Proceedings from the pediatric acute lung
Pediatric Critical Care Medicine www.pccmjournal.org      S97
Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
PALICC-2 Supplement
injury consensus conference. Pediatr Crit Care Med 2015;
16:S23–S40
2. Bernard GR, Artigas A, Brigham KL, et al: The American-
European consensus conference on ARDS. Definitions, mech-
anisms, relevant outcomes, and clinical trial coordination. Am J
Respir Crit Care Med 1994; 149:818–824
3. Force ADT, Ranieri VM, Rubenfeld GD, et al: Acute respira-
tory distress syndrome: The Berlin definition. JAMA 2012;
307:2526–2533
4. Iyer N, Khemani R, Emeriaud G, et al; Second Pediatric Acute
Lung Injury Consensus Conference (PALICC-2) for the
Pediatric Acute Lung Injury and Sepsis Investigators (PALISI)
Network: Methodology of the second pediatric acute lung in-
jury consensus conference. Pediatr Crit Care Med 2023; 24
(Suppl 1):S76–S86
5. De Luca D, Tingay DG, van Kaam AH, et al: Epidemiology of
neonatal acute respiratory distress syndrome: Prospective,
multicenter, international cohort study. Pediatr Crit Care Med
2022; 23:524–534
6. Zhu YF, Xu F, Lu XL, et al; Chinese Collaborative Study Group
for Pediatric Hypoxemic Respiratory Failure: Mortality and
morbidity of acute hypoxemic respiratory failure and acute res-
piratory distress syndrome in infants and young children. Chin
Med J (Engl) 2012; 125:2265–2271
7. Wu C, Chen X, Cai Y, et al: Risk factors associated with acute
respiratory distress syndrome and death in patients with co-
ronavirus disease 2019 pneumonia in Wuhan, China. JAMA
Intern Med 2020; 180:934–943
8. Li Y, Wang Q, Chen H, et al: Epidemiological features and
risk factor analysis of children with acute lung injury. World J
Pediatr 2012; 8:43–46
9. Killien EY, Mills B, Watson RS, et al: Morbidity and mortality
among critically injured children with acute respiratory distress
syndrome. Crit Care Med 2019; 47:e112–e119
10. Killien EY, Huijsmans RLN, Ticknor IL, et al: Acute respiratory
distress syndrome following pediatric trauma: Application of
pediatric acute lung injury consensus conference criteria. Crit
Care Med 2020; 48:e26–e33
11. Hermon M, Dotzler S, Brandt JB, et al: Extended use of the
modified Berlin definition based on age-related subgroup
analysis in pediatric ARDS. Wien Med Wochenschr 2019;
169:93–98
12. Khemani RG, Smith L, Lopez-Fernandez YM, et al; Pediatric
Acute Respiratory Distress syndrome Incidence and
Epidemiology (PARDIE) Investigators: Paediatric acute res-
piratory distress syndrome incidence and epidemiology
(PARDIE): An international, observational study. Lancet Respir
Med 2019; 7:115–128
13. Prasertsan P, Anuntaseree W, Ruangnapa K, et al: Severity
and mortality predictors of pediatric acute respiratory dis-
tress syndrome according to the pediatric acute lung injury
consensus conference definition. Pediatr Crit Care Med 2019;
20:e464–e472
14. Eworuke E, Major JM, Gilbert McClain LI: National incidence
rates for acute respiratory distress syndrome (ARDS) and
ARDS cause-specific factors in the United States (2006-
2014). J Crit Care 2018; 47:192–197
 Yehya et al
15. Parvathaneni K, Belani S, Leung D, et al: Evaluating the per-
formance of the pediatric acute lung injury consensus confer-
ence definition of acute respiratory distress syndrome. Pediatr
Crit Care Med 2017; 18:17–25
16. Yehya N, Thomas NJ, Khemani RG: Risk stratification
using oxygenation in the first 24 hours of pediatric acute
respiratory distress syndrome. Crit Care Med 2018; 46:
619–624
17. Lopez-Fernandez YM, Smith LS, Kohne JG, et al; Pediatric
Acute Respiratory Distress Syndrome Incidence and
Epidemiology (PARDIE) V3 Investigators and the Pediatric
Acute Lung Injury and Sepsis Investigators (PALISI) Network:
Prognostic relevance and inter-observer reliability of chest-
imaging in pediatric ARDS: A pediatric acute respiratory dis-
tress incidence and epidemiology (PARDIE) study. Intensive
Care Med 2020; 46:1382–1393
18. Kohne JG, Dahmer MK, Weeks HM, et al: Impact of bilat-
eral infiltrates on inflammatory biomarker levels and clinical
outcomes of children with oxygenation defect. Crit Care Med
2020; 48:e498–e504
19. Shen H, Qu D, Na W, et al: Comparison of the OI and PaO2/
FiO2 score in evaluating PARDS requiring mechanical ventila-
tion. Pediatr Pulmonol 2021; 56:1182–1188
20. Lopez-Fernandez Y, Azagra AM, de la Oliva P, et al; Pediatric
Acute Lung Injury Epidemiology and Natural History (PED-
ALIEN) Network: Pediatric acute lung injury epidemiology and
natural history study: Incidence and outcome of the acute res-
piratory distress syndrome in children. Crit Care Med 2012;
40:3238–3245
S98      www.pccmjournal.org February 2023 • Volume 24 • Number 2 (Suppl 1)
Copyright © 2023 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies.
Unauthorized reproduction of this article is prohibited
21. Yehya N, Harhay MO, Klein MJ, et al; Pediatric Acute Respiratory
Distress Syndrome Incidence and Epidemiology (PARDIE) V1
Investigators and the Pediatric Acute Lung Injury and Sepsis
Investigators (PALISI) Network: Predicting mortality in children
with pediatric acute respiratory distress syndrome: A pediatric
acute respiratory distress syndrome incidence and epidemi-
ology study. Crit Care Med 2020; 48:e514–e522
22. Slain KN, Rotta AT, Martinez-Schlurmann N, et al: Outcomes of
children with critical bronchiolitis meeting at risk for pediatric
acute respiratory distress syndrome criteria. Pediatr Crit Care
Med 2019; 20:e70–e76
23. Shein SL, Maddux AB, Klein MJ, et al; V4 PARDIE Investigators
and the PALISI Network: Epidemiology and outcomes of criti-
cally ill children at risk for pediatric acute respiratory distress syn-
drome: A pediatric acute respiratory distress syndrome incidence
and epidemiology study. Crit Care Med 2022; 50:363–374
24. Milesi C, Requirand A, Douillard A, et al: Assessment of peak
inspiratory flow in young infants with acute viral bronchiolitis:
Physiological basis for initial flow setting in patients supported
with high-flow nasal cannula. J Pediatr 2021; 231:239–245.e1
25. Matthay MA, Thompson BT, Ware LB: The Berlin definition of
acute respiratory distress syndrome: Should patients receiving
high-flow nasal oxygen be included? Lancet Respir Med 2021;
9:933–936
26. Yehya N, Khemani RG, Erickson S, et al; Pediatric Organ
Dysfunction Information Update Mandate (PODIUM)
Collaborative: Respiratory dysfunction criteria in critically ill
children: The podium consensus conference. Pediatrics 2022;
149:S48–S52