Robinder G.

Khemani MD, MsCI

Associate Professor of Clinical Pediatrics
University of Southern California Keck School of Medicine
Attending Physician, Critical Care Medicine
Children’s Hospital Los Angeles
 Paid Consultant/Advisor
 OrangeMed (Nihon Kohden)
 Hamilton Medical
 Chapter 1: Historical ARDS definitions
 Chapter 2: The Pediatric Acute Lung Injury
Consensus Conference (PALICC)
 Chapter 3: The Pediatric ARDS Incidence and
Epidemiology (PARDIE) study
 Chapter 4: Future directions
Historical Definitions of Acute Respiratory Distress Syndrome
(ARDS)
 Clinical syndrome
characterized by diffuse
inflammation in the lung
 Results in significant
hypoxemia from endothelial
injury to the pulmonary
vasculature and alveolar
epithelial injury
 No definitive diagnostic test
(i.e. biomarker) which is
present in all cases (even
histopathology
inconsistent)
Ware NEJM 2000
 Hypoxemia (Shunt)
 Poor Respiratory System Compliance
 Decreased End Expiratory Lung Volumes
(FRC)
 Diffuse Process
 Endothelial Injury
 Edema
 Increased alveolar dead space
 First description Ashbaugh -1967
 12 patients
 acute respiratory distress
 cyanosis refractory to oxygen
 decreased respiratory system compliance
 diffuse pulmonary infiltrates
 Initially Adult Respiratory Distress Syndrome
 Changed to Acute Respiratory Distress Syndrome,
to account for the fact that children have it too
First attempt at consensus definition for ARDS
Patho physiologic Concept

Diffuse vs. Lobar

Degree of Hypoxemia

Lower EELV (FRC)

Poor Compliance
Murray 1988
1. Chest roentgenogram score
No alveolar consolidation 0
Alveolar consolidation confined to 1 quadrant 1
Alveolar consolidation confined to 2 quadrant 2
Alveolar consolidation confined to 3 quadrant 3
Alveolar consolidation confined to 4 quadrant 4

2. Hypoxemia score
PaO2 / FiO2 300 0
PaO2 / FiO2 225-299 1
PaO2 / FiO2 175-224 2
PaO2 / FiO2 100-174 3
PaO2 / FiO2 <100 4

3. PEEP score cm H2O

PEEP 4 0
PEEP 5-6 1
PEEP 7-8 2
PEEP 9-11 3
PEEP 12 4

4. Respiratory system compliance score ml / cm H2O/ kg

Compliance  0.86 0
Compliance 0.75-0.85 1
Compliance 0.55 – 0.74 2
Compliance 0.30 – 0.54 3
Compliance  0.30 4
 AECC Definition AECC Limitations

Timing Acute onset  No definition of acute

 Misinterpreted as PaO2/FiO2 201-300

ALI category All patients with PaO2/FiO2<300 leading to the confusing ALI/ARDS
term
 Effect of PEEP and FiO2 on
PaO2/FiO2 ≤ 300 mmHg
Oxygenation PaO2/FiO2 ; difficulty in estimating
(regardless of PEEP)
FiO2 in non-intubated patients

Bilateral infiltrates seen on frontal  Poor interobserver reliability of

Chest Radiograph
chest radiograph chest radiograph interpretation
 High PAWP and ARDS may coexist
PAWP ≤ 18 mmHg when
 Poor interobserver reliability of
PAWP measured or no clinical evidence
of left atrial hypertension
PAWP and clinical assesments of
left atrial hypertension

Risk Factor None  Not formally included in definition

hua: CCM 2008; Wheeler: NEJM; 2006; Villar: AJRCCM 2007;Meade: AJRCCM 2000; Rubenfeld Chest 1999; Komadina: Chest 1991;
eSUPPLEMENT
1) Include radiographs of consensus interpretations for qualifying opacities vs not vs equivocal
2) Include case vignettes on how to assess “not fully explained by cardiac failure of fluid overloa
d”
 45% Mortality
Severe
14% Severe
28%
Mild
Mild
22%
22%

Moderate Moderate
64% 50%

Draft ARDS Mortality Final ARDS

PaO2/FiO2 < 100 mmHg
PEEP > 10 cm H2O
3 or 4 quadrant opacities on CXR
Crs < 40 ml/cm H2O or
VeCorr > 10 L/min
45% PaO2/FiO2 < 100 mmHg
PEEP > 5 cm H2O
Bilateral opacities on CXR
Limited Validation in Children
The Pediatric Acute Lung Injury Consensus Conference
2015
 Pediatric Adult

12.8/100,000

< 15 years

Age- and Risk-Specific Incidence of and Age-Specific

Mortality from Acute Lung Injury.
 Build off Berlin Definition- is Berlin alone Adequate?
 Pediatric Considerations
1. Timing (similar to adults?)
2. Age?
3. Co-existence with Cardiac Disease (similar to adults?)
4. What are best respiratory criteria for risk
stratification/disease severity
5. How do we handle patients on non-invasive ventilation
and/or those without arterial lines
6. How important/reliable are Radiographic Criteria for
Definition of ARDS
7. Defining it in patients with cardiac or pulmonary co-
morbidites
 Age Exclude patients with peri-natal related lung disease
Timing Within 7 days of known clinical insult
Origin of Edema Respiratory failure not fully explained by cardiac failure or fluid overload
Chest Imaging Chest imaging findings of new infiltrate(s) consistent with acute pulmonary
parenchymal disease
Non Invasive mechanical ventilation Invasive mechanical ventilation
PARDS (No severity stratification) Mild Moderate Severe
Oxygenation Full face-mask bi-level ventilation or 4 ≤ OI < 8 8 ≤ OI < 16 OI ≥ 16
2
CPAP ≥5 cm H20
PF ratio ≤ 300 5 ≤ OSI < 7.51 7.5 ≤ OSI < 12.31 OSI ≥ 12.31
SF ratio ≤ 264 1
Special Populations
Cyanotic Heart Standard Criteria above for age, timing, origin of edema and chest imaging with an
Disease acute deterioration in oxygenation not explained by underlying cardiac disease. 3
Chronic Lung Standard Criteria above for age, timing, and origin of edema with chest imaging
Disease consistent with new infiltrate and acute deterioration in oxygenation from baseline
which meet oxygenation criteria above.3
Left Ventricular Standard Criteria for age, timing and origin of edema with chest imaging changes
dysfunction consistent with new infiltrate and acute deterioration in oxygenation which meet
criteria above not explained by left ventricular dysfunction.
OI = oxygenation index = (FiO2* mean airway pressure*100)/ PaO2
OSI = oxygen saturation index = (FiO2* mean airway pressure*100) /SpO2
PALICC PCCM 2015
 Build off Berlin Definition- is Berlin alone Adequate?
 Pediatric Considerations
1. Timing (similar to adults?)
2. Age?
3. Co-existence with Cardiac Disease (similar to adults?)
4. What are best respiratory criteria for risk
stratification/disease severity
5. How do we handle patients on non-invasive ventilation
and/or those without arterial lines
6. How important/reliable are Radiographic Criteria for
Definition of ARDS
7. Defining it in patients with cardiac or pulmonary co-
morbidites
1.5.1 Oxygenation Index [OI=(FiO2* mean airway
pressure*100)/ PaO2], in preference to PaO2/FiO2
(PF) ratio, should be the primary metric of lung
disease severity to define P-ARDS for all patients
treated with invasive mechanical ventilation.
Strong agreement
1.5.2 PaO2/FiO2 (PF) ratio should be used to diagnose
ARDS for patients receiving non‐invasive full face
mask ventilation (CPAP or BiPAP) with a
minimum CPAP of 5 cm H2O. Strong agreement
PALICC PCCM 2015
 Lower and more variable ventilator
management and in particular PEEP use in
Pediatrics compared to Adults
 Need to control for ventilator support
 More Frequent use of HFOV

PALICC PCCM 2015

 OI versus PF ratio

PEEP 5 cmH20 PEEP 14 cmH20

PaO2 = 60, FiO2=0.6, MAP = 8 PaO2 = 120, FiO2=0.5, MAP = 19
PF Ratio 100 PF Ratio 240
OI =8 OI = 8
 Study PEEP
Flori 2005 5.3 +/- 2.6
Erikson 2007 8.5 (7,11)
Khemani 2009 6 (4,8)
Santishi 2010 (PALIVE)* 6.9 +/- 2.7
Lopez-Fernandez 2012 8.9 +/- 2.9
* > 50% of patients with PEEP ≤5 cmH20

> 50% of adults in ARDSNet studies on PEEP

≥10cmH20 at study inclusion (Britos CCM)
 Newth CPPCRN 2014

Khemani ICM 2011

Because PEEP management is not

protocolized in children, it may
confound ascertainment of disease
severity (alters PF ratio) and ability to
examine other treatment effects (Tidal
Volume etc.)
Santichi PALIVE 2011
PALICC PCCM 2015
 OI >16 8-16 4-8 <4 Total
Category Severe Moderate Mild At Risk
Derivation Set
Khemani 2009
Number 98 (24.7%) 104 (26.2%) 147 (37.1%) 48 (12.1%) 397
Mortality 40 (40.8%) 21 (20.2%) 23 (11.6%) 2 (4.2%) 80 (20.2%)
Validation Set
Flori 2005
Number 28 (16.4%) 60 (35.1%) 67 (39.2%) 16 (9.4%) 171
Mortality 10 (35.5%) 16 (26.7%) 13 (19.4%) 1 (6.25%) 40 (23.4%)
Curley 2005
Number 34 (40%) 28 (33%) 20 (24%) 3 (3%) 85
Mortality 2 (5.9%) 4 (14.3%) 1 (5%) 0 (0%) 7 (8.2%)
Erickson 2007
Number 38 (33.3%) 31 (27.2%) 36 (31.6%) 9 (7.9%) 114
Mortality 20 (52.6%) 10 (32.3 %) 5 (13.9%) 0 (0%) 35 (30.7%)
Kneyber 2008
Number 8 (27.6%) 11 (37.9%) 10 (34.5%) 0 (0%) 29
Mortality 1 (12.5%) 3 (27.3%) 1 (10%) 0 (0%) 5 (17.3%)
Lopez-Fernandez 2012
Number 72 (55%) 47 (35.9%) 12 (9.1%) 0 (0%) 131
Mortality 24 (33.3%) 10 (21.3%) 3 (25%) 0(0%) 37 (28.2%)
Sapru 2013
Number 47 (28%) 68 (40.2%) 40 (23.7%) 14 (8.3%) 169
Mortality 9 (19.1%) 11 (16.2%) 0 (0%) 0 (0%) 20 (11.9%)
Validation Set Total
Number 225 (32.7%) 241 (34.8%) 184 (26.6%) 42 (6.1%) 692
Mortality 66 (29.3%) 54 (22.4%) 23 (12.5%) 1 (2.4%) 144 (20.8%) PALICC PCCM 2015
 Build off Berlin Definition- is Berlin alone Adequate?
 Pediatric Considerations
1. Timing (similar to adults?)
2. Age?
3. Co-existence with Cardiac Disease (similar to adults?)
4. What are best respiratory criteria for risk
stratification/disease severity
5. How do we handle patients on non-invasive ventilation
and/or those without arterial lines
6. How important/reliable are Radiographic Criteria for
Definition of ARDS
7. Defining it in patients with cardiac or pulmonary co-
morbidites
1.6.1 Oxygen Saturation Index [OSI =(FiO2* mean
airway pressure*100)/SpO2]should be used
when an OI is not available for stratification
of risk for patients receiving invasive
mechanical ventilation. Strong agreement
1.6.2 SF (SpO2/FiO2)ratio can be used when PF
ratio is not available to diagnose P-ARDS in
patients receiving non‐invasive full face
mask ventilation (CPAP or BiPAP) with a
minimum CPAP of 5 cm H2O. Strong
agreement
1.8.1 To apply SpO2
criteria to
diagnose P-ARDS,
oxygen therapy
should be titrated
to achieve an SpO2
between 88- 97%.
Strong
agreement
PALICC PCCM 2015
Khemani CCM 2012
Parvathaneni PCCM 2016
Requiring PF ratio selects for patients with hypoxemia and cardiovascular dysfunction

Not using pulse oximetry criteria underestimates the prevalence

of ARDS and over-estimates mortality, by selecting for a more ill
patient population
 Build off Berlin Definition- is Berlin alone Adequate?
 Pediatric Considerations
1. Timing (similar to adults?)
2. Age?
3. Co-existence with Cardiac Disease (similar to adults?)
4. What are best respiratory criteria for risk
stratification/disease severity
5. How do we handle patients on non-invasive ventilation
and/or those without arterial lines
6. How important/reliable are Radiographic Criteria for
Definition of ARDS
7. Defining it in patients with cardiac or pulmonary co-
morbidites
1.4.1 Chest imaging findings of new infiltrate(s)
consistent with acute pulmonary
parenchymal disease are necessary to
diagnose P-ARDS. Strong agreement

PALICC PCCM 2015

 Eliminated requirement for bilateral infiltrates
 1. Poor Inter-observer reliability, unclear if can be
improved with common training
 2.Poor Sensitivity of CXR to detect disease, CXR lags
behind hypoxemia, poor correlation with CT
 3. Pathophysiology may be present with radiographic
unilateral disease (inability to detect inflammation in
other lung)
 4. Unclear if Bilateral Infiltrates adds anything
prognostically over hypoxemia

PALICC PCCM 2015

 Two adult studies demonstrating no prognostic
relevance when controlling for hypoxemia, one
showing important (Luhr 1999,2000, Roupie 1999)
 Two pediatric studies in China showing prognostic
importance, but not adequately controlling for
hypoxemia (ALI v ARDS) ( Hu 2010, Zhu 2012)
 Secondary Analysis of Published data (CHLA cohort,
ICM 2009)
PF Ratio <100 100-200 200-300 Total
Number of Patients 134 (33.8%) 177 (44.6%) 86 (21.6%) 397
Bilateral Infiltrates 82 (61%) 71 (40.1%) 39 (45.3%) 192 (48.4%)
Mortality with Bilat Inf 29 (35.4%) 10 (14.1%) 5 (12.8%) 44 (22.9%)
Mortality without Bilat Inf 19 (36.5%) 14 (13.2%) 3 (6.4%)* 36 (17.6%)
PALICC 2015
The Pediatric ARDS Incidence and Epidemiology Study (PARDIE)
 PARDIE
The Pediatric Acute Respiratory Distress syndrome Incidence
and Epidemiology
Executive committee: study
Philippe Jouvet, Robinder Khemani,
Christopher Newth, Lincoln Smith, Neal Thomas, Doug
Willson
PARDIE is an international initiative to better understand the implications of the
new Pediatric Acute Lung Injury of
On Behalf Consensus
the PARDIEConference (PALICC) definition of
Investigators
Pediatric ARDS

http://pardie.palisi.org/
 CHLA, RSR, St Justine
▪ PARDIE Data entry Website/Database, Admin Support
 SC CTSI
▪ To support secondary analysis (statistical support) and data
management
 International cross sectional (incidence) design to
test several aspects of the new definition
 Designed as a Main study (V.0) plus ancillaries
 5 continuous days of screening and patient enrollment (M-
F), performed 10 times in a year (50 total study days)
 Only new cases of ARDS diagnosed within previous 24
hours included
 Data collection largely in first 3 days of ARDS diagnosis
 Follow clinical outcomes such as mortality and length of
ventilation for included patients.
 1. To determine the number and frequency of new cases of
PARDS amongst non-cardiac PICU patients, and their
respective outcomes, including ICU and hospital mortality
and VFDs
 2. To evaluate how the PALICC recommended mild,
moderate, and severe classification of PARDS performs in
discriminating ICU and hospital mortality
 3. To determine the relevance of the timing of hypoxemia
metrics (OI, OSI, PF, SF) within the first three days of
ARDS onset affects the discrimination ability of ICU and
hospital mortality
  SpO2 criteria
 Simplified CXR
 CHD and CLD pts
 Co-existence with
cardiac disease
 Explicit criteria for NIV
Parvatheneni PCCM 2017
 1. To determine the number and frequency of new cases of
PARDS amongst non-cardiac PICU patients, and their
respective outcomes, including ICU and hospital mortality
and VFDs
 2. To evaluate how the PALICC recommended mild,
moderate, and severe classification of PARDS performs in
discriminating ICU and hospital mortality
 3. To determine the relevance of the timing of hypoxemia
metrics (OI, OSI, PF, SF) within the first three days of
ARDS onset affects the discrimination ability of ICU and
hospital mortality
 OI >16 8-16 4-8 <4 Total
Category Severe Moderate Mild At Risk
Derivation Set
Khemani 2009
Number 98 (24.7%) 104 (26.2%) 147 (37.1%) 48 (12.1%) 397
Mortality 40 (40.8%) 21 (20.2%) 23 (11.6%) 2 (4.2%) 80 (20.2%)
Validation Set
Flori 2005
Number 28 (16.4%) 60 (35.1%) 67 (39.2%) 16 (9.4%) 171
Mortality 10 (35.5%) 16 (26.7%) 13 (19.4%) 1 (6.25%) 40 (23.4%)
Curley 2005
Number 34 (40%) 28 (33%) 20 (24%) 3 (3%) 85
Mortality 2 (5.9%) 4 (14.3%) 1 (5%) 0 (0%) 7 (8.2%)
Erickson 2007
Number 38 (33.3%) 31 (27.2%) 36 (31.6%) 9 (7.9%) 114
Mortality 20 (52.6%) 10 (32.3 %) 5 (13.9%) 0 (0%) 35 (30.7%)
Combination of studies, several Kneyber 2008

years old Number

Mortality
8 (27.6%)
1 (12.5%)
11 (37.9%)
3 (27.3%)
10 (34.5%)
1 (10%)
0 (0%)
0 (0%)
29
5 (17.3%)
Lopez-Fernandez 2012
Number 72 (55%) 47 (35.9%) 12 (9.1%) 0 (0%) 131
No use of Pulse oximetry Criteria Mortality 24 (33.3%) 10 (21.3%) 3 (25%) 0(0%) 37 (28.2%)
Sapru 2013
Number 47 (28%) 68 (40.2%) 40 (23.7%) 14 (8.3%) 169
Is this representative of current Mortality 9 (19.1%) 11 (16.2%) 0 (0%) 0 (0%) 20 (11.9%)

practice and outcomes? Number 225 (32.7%)

Validation Set Total
241 (34.8%) 184 (26.6%) 42 (6.1%) 692
Mortality 66 (29.3%) 54 (22.4%) 23 (12.5%) 1 (2.4%) 144 (20.8%)
 1. To determine the number and frequency of new cases of
PARDS amongst non-cardiac PICU patients, and their
respective outcomes, including ICU and hospital mortality
and VFDs
 2. To evaluate how the PALICC recommended mild,
moderate, and severe classification of PARDS performs in
discriminating ICU and hospital mortality
 3. To determine the relevance of the timing of hypoxemia
metrics (OI, OSI, PF, SF) within the first three days of
ARDS onset affects the discrimination ability of ICU and
hospital mortality
Initial versus a Delayed Value?
OI/OSI
Only OI

Yehya, PCCM 2015 Parvatheneni, PCCM 2017

Perhaps 12 hours post PARDS captures risk well?
 V.0 – PARDIE
 V.1 – Risk factors for Mortality in PARDS
 V.2 – Monitoring/Ventilator Management in PARDS
 V.3 – Chest Imaging

 V.4 – At Risk for ARDS

 Cardiac PARDIE
 97 is the new 100
 May 9-13, 2016 (103 sites)
 July 11-15, 2016 (131 sites)
 September 12-16, 2016 (141 sites)
 November 14-18, 2016 (165 sites)
 January 9-13, 2017 (166 sites)
 February 13-17,2017 (129 sites)
 March 20-24, 2017 (128 sites)
 April 17-21, 2017 (129 sites)
 May 15-19, 2017 (128 sites)
 June 12-16, 2017 (131 sites)
 PARDIE
THE PEDIATRIC ACUTE RESPIRATORY DISTRESS
SYNDROME INCIDENCE AND EPIDEMIOLOGY STUDY

50
45 9

38 3
 Country Region Income Classification

Argentina Central & South America Middle-High

Australia Middle East/Asia/Australia/New Zealand High
Bolivia Central & South America Middle-Low*
Canada North America High
Chile Central & South America High
China Middle East/Asia/Australia/New Zealand Middle-High
Colombia Central & South America Middle-High
Czech Republic Europe Middle-High
Ecuador Central & South America Middle-High
France Europe High
Greece Europe High
Italy Europe High
Japan Middle East/Asia/Australia/New Zealand High
Malaysia Middle East/Asia/Australia/New Zealand Middle-High
Mexico Central & South America Middle-High
Lancet Respiratory Medicine
Netherlands Europe High
New Zealand Middle East/Asia/Australia/New Zealand High
145Panama
International Sites
Central & South America Middle-High
Peru Central & South America Middle-High
289 Investigators
Poland Europe High
Portugal Europe High
27 Countries
Saudi Arabia Middle East/Asia/Australia/New Zealand High
Singapore Middle East/Asia/Australia/New Zealand High
Spain Europe High
Turkey Middle East/Asia/Australia/New Zealand Middle-High
UK Europe High
PARDIE, Lancet Respiratory Medicine
 Lancet Respiratory Medicine 2018

 PALICC PARDS Incidence and outcomes

 PALICC vs. Berlin comparisons
 Pulse Oximetry
 Bilateral Infiltrates
 PF vs OI
 Hypoxemia severity, timing with mortality
 NIV PARDS descriptive analysis
 Multivariable model for mortality with focus
on PARDS definition
 145 PICUs
 23,315 patients admitted
 12,270 were on study defined mechanical
ventilation
 744 new cases of PARDS

 3.1% amongst ICU patients

 6.1% amongst those on MV

PARDIE, Lancet Respiratory Medicine

  PARDS All Patients MV Patients
  N N Incidence (95% C.I.) P value N Incidence (95% C.I.) P value
All Patients 744 23280 3.20% (3.0, 3.4) NA 12242 6.08% (5.7, 6.5) NA
Season       <0.0001     0.0002
Fall 156 4942 3.16% (2.7, 3.6)   2565 6.08% (5.2, 7.0)  
Winter 286 7479 3.82% (3.4, 4.3)   4121 6.94% (6.2, 7.7)  
Spring 210 6382 3.29% (2.9, 3.7)   3327 6.31% (5.5, 7.1)  
Summer 92 4477 2.05% (1.6, 2.5)   2229 4.13% (3.3, 5.0)  
Region       0.002     <0.0001
North America 451 12843 3.51% (3.2, 3.8)   6023 7.49% (6.8, 8.2)  
Europe 127 4761 2.67% (2.2, 3.1)   2743 4.63% (3.8, 5.4)  
Central/South America 126 3836 3.28% (2.7, 3.8)   2389 5.27% (4.4, 6.2)  
Middle
40 1840 2.17% (1.5, 2.8)   1087 3.68% (2.6, 4.8)  
East/Asia/Australia/NZ
Country Income       0.23     0.0002
High Income 617 18905 3.26% (3.0, 3.5)   9446 6.53% (6.0, 7.0)  
Middle/Low Income 127 4375 2.90% (2.4, 3.4)   2796 4.54% (3.8, 5.3)  

PARDIE, Lancet Respiratory Medicine

PARDIE, Lancet Respiratory Medicine
PARDIE, Lancet Respiratory Medici
 Mild PARDS Moderate PARDS Severe PARDS p value

210 145 165

PICU Mortality 26 (12.4%) 15 (10.3%) 54 (32.7%) <0.001
90 Day Mortality & 30 (14.3%) 21 (14.5%) 56 (34%) <0.001
28 Day VFDS IMV Only 21.1 (12.9,23.9) 20.5 (13.2,23.9) 13.1 (0,21.0)* # <0.001
28 Day VFDS IMV or NIV 20.5 (10.6,23.6) 18.4 (8.9,23.4) 10.1 (0,20.4)* # <0.001

Length IMV Survivors 5.9 (4,10.2) 7.0 (4.0,12.4) 7.9 (5.0,14.7)* 0.01
Length IMV/NIMV 6.5 (4.1,11.6) 7.47 (4.0,14.6) 9.25 (5.9,17.1)* 0.004
Survivors

• * indicates different than mild PARDS in multiple comparisons

• # indicates different than moderate PARDS in multiple comparisons.

PARDIE, Lancet Respiratory Medicine

PARDIE, Lancet Respiratory Medicine
PARDIE, Lancet Respiratory Medicine
 Time PALICC Berlin
N AUC AUC P
PARDS 518 0.644 0.605 0.09
Onset (0.582,0.705) (0.548,0.663)
6 Hours 437 0.689 0.639 0.01
(0.624,0.754) (0.575,0.703)
12 Hours 427 0.683 0.597 0.0001
(0.621,0.745) (0.529,0.666)
18 Hours 420 0.670 0.613 0.02
(0.605,0.736) (0.548,0.679)
24 Hours 418 0.675 0.618 0.02
(0.610,0.740) (0.547,0.689)
30 Hours 412 0.685 0.648 0.09
(0.621,0.749) (0.580,0.717)
36 Hours 400 0.659 0.643 0.45
(0.581,0.737) (0.573,0.714)
42 Hours 387 0.641 0.628 0.58
(0.565,0.717) (0.557,0.699)
48 Hours 381 0.682 0.639 0.05
(0.610,0.754) (0.562,0.716)
   NIV PARDS NIVPARDS NIV PARDS Not  
Intubated Intubated
  160 80 80  
PICU Mortality 24 (15%) 20 (25%) 4 (5%) <0.001
90 Day Mortality & 28 (17.5%) 22 (27.5%) 6 (7.5%) <0.001
28 Day VFDS IMV Only 25.3 (14.2,28) 17.5 (0,23.1) 28 (28,28) <0.001
28 Day VFDS IMV or NIV 20.4 (3.3,24.9) 14.4 (0, 21) 24.5 (20.1,25.9) <0.001

Length IMV Survivors 0 (0,6.6) 7.6 (4.3,12.5) NA  

Length IMV/NIMV 5.9 (2.7,10.7) 10.7 (5.7,20.5) 3.2 (2.0,6.9) <0.001
Survivors

½ of all NIV PARDS patients are intubated, 87% within 48 Hours of PARDS

NIV PARDS patients who are intubated have mortality between moderate and severe

PARDIE, Lancet Respiratory Medicine

PARDIE, Lancet Respiratory Medicine
 Site PI: Bilateral infiltrates
 74% of children at PARDS diagnosis
 87% within 3 days of PARDS diagnosis
 759 chest x-rays evaluated by both a pediatric
intensivist and a radiologist (v.3)
 poor agreement regarding bilateral infiltrates
(kappa= 0.34).

PARDIE, Lancet Respiratory Medicine

Risk factors for Mortality OR (95% CI) P value

Severe PARDS at onset 3.04 (1.89, 4.87) <0.001

Medium Income Country 2.85 (1.56, 5.22) 0.001
PARDS Risk Factor 0.48 (0.30, 0.76) 0.002
Pneumonia
PARDS Risk Factor Other 10.9 (2.56, 46.5) 0.001
Shock
PARDS Risk Factor Drowning 12.7 (2.0, 79.4) 0.007

Immune Suppression 6.3 (3.6, 11.2) <0.001

* Bilateral infiltrates not significant p>0.2
PARDIE, Lancet Respiratory Medicine
 PARDS affects 3% of PICU patients
 Mortality > 30% for severe hypoxemia
 PALICC identified 40% more children as having ARDS compared
to Berlin (with arterial PF ratio), and diagnoses ARDS a median of
12-30 hours sooner.
 PALICC based OI/OSI severity groups appear to stratify mortality
better than PF/SF groups
 PARDS severity 6 hours after PARDS diagnosis may be good
marker for risk
 PALICC groups are calibrated for VFD and length of MV in
survivors.
 Many factors outside of PARDS severity associated with mortality

PARDIE, Lancet Respiratory Medicine

Parameter Target Range Notes
Tidal Volume 5-8 ml/kg PBW 3-6 ml/kg severe PARDS
Plateau Pressure < 28 cmH20 < 32 high CW elastance
PEEP 10-15 cm H20 moderate to Higher as necessary,
severe PARDS monitoring compliance,
hemodynamics, oxygen
delivery
pH 7.15-7.30 No clear lower limit
Oxygenation SpO2 92-97 PEEP < 10 No clear lower limit
SpO2 88-92% PEEP ≥ 10
 Lancet Respiratory Medicine 2018

 PARDS Severity
  All (n=3818) No PARDS Mild Moderate Severe
(n=1012) (n=1229) (n=1075) (n=502)
VT (ml/kg) ABW 6.7 (5.7, 8) 6.7 (5.7, 7.8)  6.8 (5.8, 8)  6.6 (5.6, 8)  6.6 (5.4, 7.8) 
∆Pressure (PIP- 18 (14, 22) 14 (11, 18) 18 (14, 21) 19 (15, 23) 21 (17, 26)
PEEP)

PIP 26 (21, 30) 21 (18, 25) 25 (22, 29) 28 (24, 32) 32 (28, 36)
PEEP 8 (6, 10) 6 (5.5, 8) 8 (6, 9) 9 (7, 10) 10 (8, 12)
PaCO2 44 (38, 53) 41 (36, 47) 43 (38, 51) 46.5 (39, 56) 48 (41, 60)
pH (ABG only) 7.38 (7.3, 7.4) 7.41 (7.3, 7.4) 7.39 (7.3, 7.4) 7.37 (7.3, 7.4) 7.32 (7.2, 7.3)
N = 600 PARDS Patients, 100 International ICUs, q 6 hr.
 V.1. Risk factors for mortality and practice patterns for
ancillary therapies
 V.2 Ventilator management and monitoring practices and
relationship with outcome
 V.3 Chest imaging and inter-observer variability,
relationship with outcome, automated methodologies
 V.4. At risk for PARDS group and outcomes, prediction of
PARDS progression, feasibility of PARDS prevention
studies
 97 is the new 100- QI paper describing methodology for
decreasing FiO2
 Public Use dataset: An anonymous version of
the PARDIE dataset will be made publicly
available approximately five (5) years after
publication of the Main PARDIE Paper, with
the relevant case report forms to facilitate
data analysis.
What is next?
High level of evidence lacking in nearly all PARDS recommendations
 16 current clinical trials in pediatric ARDS on
clinicaltrials.gov
 Variety of multi-center clinical trials in
planning stages through international
organizations
 Secondary analysis of large datasets of
PARDS cohorts
Have we generated enough new evidence?