Pediatric acute respiratory distress syndrome (PARDS) is common in PICUs around the world. Nearly 1 out of every 100 children admitted to the PICU meet the American European Consensus Conference criteria (1) for acute lung injury (ALI) or acute respiratory distress syndrome (ARDS), and pediatric-specific studies demonstrate the incidence to be from 1.4 cases/100,000/yr to 9.5 cases/100,000/yr (2–7). While the pooled mortality from several studies is about 30%, it is clear that children with certain comorbidities have worse outcomes. This article will discuss the impact of these comorbidities on the severity of PARDS.
Comorbidities and PARDS
The comorbidities associated with outcome in pediatric patients with PARDS can be divided into 1) those factors that are patient specific and 2) those factors inherent to the PARDS disease process. Severity of disease is often described with regard to factors identifiable at the time the patient arrives in the ICU, that is, those factors resultant from the length of time to diagnosis and early management of the patient. The reader is encouraged to review the article on pathobiology in this journal edition for in-depth discussion of the mechanisms of PARDS (8). Importantly, comorbidities and severity of disease may not be mutually exclusive. Many measures that are predictive and made once the patient is diagnosed, such as PaO2/FIO2 (P/F) ratio, are likely influenced by the underlying disease process itself, such as pneumonia or degree of smoke inhalation, but may also be influenced by the patient’s underlying genetic predisposition to lung injury, nutritional status, or chronic comorbidity. For simplicity, however, in the following text, comorbidities and severity of illness will be handled separately.
The overall mortality associated with pediatric ALI is decreasing with time (2) (Fig. 1). Most recent data indicate a mortality of 18–22% for both invasively and noninvasively ventilated patients with ALI (2, 9). Of the studies examining invasively ventilated patients, mortalities are somewhat higher, 26–35% (5, 7, 10, 11).
Comorbidities: Patient-Specific Factors
Although patients with underlying chronic disease constitute a significant proportion of children with PARDS, most studies do not indicate that chronic disease, other than immunodeficiency, contribute to an increased risk of death or prolonged mechanical ventilation independent of other important covariates such as multiple organ system dysfunction and severity of lung disease at the onset of PARDS.
The primary comorbidity associated with poor outcome from PARDS is preexisting immunodeficiency, including both congenital and acquired immunodeficiencies. Most studies prior to 2000 report mortalities around 70% for patients with ALI/ARDS and HIV and 84–95% for those with malignancy or after bone marrow transplant who develop ALI/ARDS (12–17). As many immunodeficient patients manifest severe oxygenation defect, sepsis, and/or multiple organ system failure, presence of immunodeficiency often “falls out” of multivariate analyses, particularly in smaller studies (15–18), although it is still an independent risk factor in other studies (13).
Data published after 2000 indicate that mortality in this cohort has been decreasing, with many studies reporting mortality in the 70% range (5, 19, 20). Nonetheless, compared to an overall pediatric ALI mortality of approximately 18–35%, this group reflects a disproportionately high risk of death, with the effect sufficiently strong to have altered the outcome of at least one pediatric interventional trial (21). Although most oncologic patients with immunodeficiency present with neutropenia, it must be noted that the severity of illness of these patients has been reported to worsen as neutrophil count recovers (22).
Gender, Race, and Ethnicity.
In 2009, the ARDS Network published a large observational study of 2,362 mechanically ventilated adults with ALI. After adjusting for demographic and clinical covariates, African-American and Hispanic adults exhibited an increased risk of death and fewer ventilator-free days, although the increased mortality in African-Americans appeared to be related to the higher severity of illness at the time of presentation. The increased risk of mortality in Hispanics was independent of multiple covariates including severity of illness (23). Unlike the adult data (24–26) that indicate no association between gender and mortality, a pediatric study published in 1992 reported a higher mortality for female patients (82% vs 62% male patients; p = 0.04) that persisted on multivariate analysis (p = 0.038) in a cohort of 100 patients (13). Male patients were reported to have an independently increased risk of death in a cohort of 39 pediatric patients with respiratory failure after bone marrow transplantation (14). Although recent pediatric data on ventilator-associated pneumonia also seem to show some signal toward worse outcome in female patients on multivariable analysis (27), many more recent pediatric ALI/ARDS investigations have shown no directionality with regard to gender or ethnicity (2, 5, 9–11, 15, 18).
Age does not appear to be associated with clinical outcomes such as mortality or length of mechanical ventilation in most studies of children with ALI/ARDS (2, 9–11, 13, 18). However, a lower mortality in younger children on both univariate (median age of survivors, 1.0 yr; interquartile range [IQR], 0.4–3.1 yr vs nonsurvivors, 4.9 yr; IQR, 0.8–9.8 yr) and multivariate analysis after controlling for severity of illness (Pediatric Index of Mortality [PIM]-2), etiology, immunodeficiency, and multiple respiratory indices has been reported (p = 0.027) (5). By contrast, in a study of 345 Chinese children with acute hypoxemic respiratory failure, infants less than 1 year old had a higher mortality (odds ratio [OR], 2.1; 95% CI, 1.17–3.49) and fewer ventilator-free days (OR, 2.65; 95% CI, 1.57–4.47) (28).
2.1.1 There may be a difference in the progression and outcome from ARDS in children when compared with adults. We recommend that future studies be designed to examine whether there are differences in the progression and/or outcome of ARDS between adults and children or between children of different ages. Strong agreement
Several studies in adults and children have demonstrated associations between genetic variations in genes coding for proteins involved in the pathophysiology of ALI/ARDS and either the development or severity of lung injury. Adult studies have examined such genes as angiotensin-converting enzyme (29–31), angiopoietin-2 (32, 33), Factor V Leiden (34), hemeoxygenase 2 (35), interleukin (IL)-1 receptor antagonist (36), IL-6 (37–40), mannose-binding lectin 2 (41, 42), myosin light chain kinase (43, 44), plasminogen activator inhibitor-1 (45), pre-B-cell colony-enhancing factor (46), protein tyrosine phosphatase, receptor type, f polypeptide, interacting protein coding for liprin α (47), toll-like receptor 1 (48), tumor necrosis factor α (49), and vascular endothelial growth factor (50, 51). Although there are few comparable studies in the pediatric population, a few studies suggest some degree of influence of host genetic variation with the development and severity of ALI/ARDS in children (52–54).
2.1.2 There is a paucity of studies related to the pathophysiology of PARDS. The impact of postnatal maturational development on the pathophysiology of PARDS is unknown. We recommend that biomarker and genetic studies that may provide insight into the pathophysiology of PARDS in children, and study of pathophysiology in animals of different ages with age cutoffs informed by chronology of postnatal lung and immune system development, should be a focus of future research protocols. Strong agreement
Comorbidities: Disease-Specific Factors
The most common disease process associated with pediatric ALI/ARDS is pneumonia (9), as it is in the adult population (26) with sepsis often listed as the second major cause (Fig. 2). Univariate analyses often report a lower risk of all-cause mortality and/or prolonged mechanical ventilation in those patients with “direct” causes of ALI, such as pneumonia and aspiration, compared with patients with “indirect” causes of ALI, such as sepsis (5, 7, 9, 55). However, caution is warranted as mortality in the cohorts with “direct” causes of ALI (31.4% , 23.5% , 11% ) is generally still much greater than the average mortality of all PICU patients. Also, some infectious processes with a predilection for primary pneumonia are inherently more virulent within this category and include pandemic H1N1 (56), severe acute respiratory syndrome (57), and most recently, methicillin-resistant Staphylococcus aureus and methicillin-sensitive S. aureus infection, particularly when there is coinfection with influenza (58).
For the primary cause of “indirect” ALI, sepsis, assessment of the literature is confounded because some studies examine only risks of sepsis and their association with clinical outcome, whereas others only include patients with overt septic shock. That said, although overall mortality for pediatric sepsis and septic shock (with and without ALI) is approximately 10%, it is in those patients with sepsis with ALI, with mortality ranging from 17%, as reported in the REsearching severe Sepsis and Organ dysfunction in children: a gLobal perspectiVE trial of drotrecogin-alpha in pediatric severe sepsis up to 70% in various studies (5, 9, 16, 18, 59). Even in the most conservative of these estimates, a significantly higher risk of death (OR, 3.55; 95% CI, 1.48–8.50; p = 0.004) and fewer ventilator-free days (OR, 2.11; 95% CI, 1.02–4.36; p = 0.04) in sepsis patients with ALI has been reported (9). For many studies, however, the diagnosis associated with ALI/ARDS often “falls out” on multivariate analysis, likely a result of small study size (7, 9, 18, 60).
Although not universal, some studies report persistence of sepsis as an independent risk factor associated with mortality on multivariate analysis. Additional studies are necessary, either larger prospective observational studies (on the international level) or meta-analysis of multiple studies, provided that the diagnosis of nonpulmonary sepsis is consistently defined across all the centers and studies being included in the investigation.
ASSESSING DISEASE SEVERITY
Assessing the severity of disease in patients with PARDS has been much more widely studied in adults than in children. Disease severity measures can be subdivided into 1) measures that can be made at the bedside, 2) measures requiring more in-depth calculation, 3) biochemical measurements, and 4) early responsiveness to therapy.
Measures That Can Be Made at the Bedside
Despite inconsistency in the optimal timing of the measurement, both oxygenation defect and ventilation defect have been demonstrated in multiple studies to be associated with outcome. Some studies report measurements of oxygenation and ventilation defects at the onset of ALI/ARDS, whether the patient is invasively or noninvasively ventilated at the time. By contrast, other studies report these measurements within the first hours to days of ALI/ARDS onset, whereas others report the association of these measures with outcome based on the “worst” measurement during the patient’s entire PICU stay.
Measurements of Oxygenation Indices.
Of the oxygenation measures, both P/F ratio and oxygenation index (OI) have been fairly consistent and robust predictors of disease severity and outcomes (5, 7, 9, 28, 61). Given the decrease in use of arterial catheters and, hence, arterial blood gases in the pediatric population and the increased use of noninvasive respiratory support strategies, SpO2/FIO2 and oxygen saturation index (OSI) have been shown to be valid substitutions for the arterial blood gas–based measurements (61–63). These findings are critically important in allowing earlier identification of patients for clinical purposes as well as appropriate identification of patients for inclusion in future PARDS research. The reader is encouraged to review the article on definitions, incidence, and epidemiology for additional details on oxygenation defect measurement and defining PARDS severity (64).
Timing of Measuring Oxygenation Indices.
When assessing the timing of these measurements in the course of PARDS, strong predictive ability in measurements made at the onset of or perhaps more accurately at the time that PARDS is diagnosed has been reported. Specifically, the P/F ratio at the onset of ALI/ARDS has been shown to be an independent predictor of mortality and duration of mechanical ventilation (9). This finding was confirmed in a more recent pediatric cohort (7). Certainly, this finding is multifactorial and may be related to earlier diagnosis of the patient and institution of lung-protective management strategies rather than the measure itself being causative of the outcome.
The utility of measurements reflecting persistence of oxygenation defect after the onset (or diagnosis) of ALI/ARDS in predicting outcome has been well demonstrated (5, 7, 10, 11, 18, 61). Children with persistently elevated P(A-a)O2 greater than 420 mm Hg from day 2 to day 5 after diagnosis demonstrate an increased risk of death with an OR of 26.7 (95% CI, 4.7–183) in a study of 60 patients with ARDS (17). P/F ratio after 24 hours of ARDS onset was able to predict mortality (p < 0.001) in a cohort of children stratified into four categories (severe ARDS, moderate ARDS, ALI, and acute respiratory failure) (7) (Fig. 3). Increased mortality has been observed in children with lower P/F ratio after 24 hours of ALI/ARDS (5, 10). In addition, both maximum oxygen index (OI) and minimum P/F ratio are good predictors of mortality; a minimum P/F ratio of less than 53 predicted risk of death greater than 70.5% with specificity of 92% and maximum OI greater than 53.5 predicted risk of death greater than 80% with specificity of 97% (5). Lower P/F ratio during hospital stay has also been observed to predict mortality (60).
Measurements of Ventilation Indices.
Oxygenation reflects only a portion of the physiological deficits in PARDS. Recent pediatric studies have demonstrated the predictive value of CO2 exchange, and thereby lung perfusion, on outcomes just as has been seen in the adult population (11, 15, 16, 60, 64, 66). Worse lung perfusion is thought to be due to dysregulated coagulation and vascular endothelial injury in the lungs of PARDS patients. Abnormal ventilation indices (VI = PaCO2 × peak airway pressure × respiratory rate/1,000) at the onset of disease in children with ARDS have been reported. Furthermore, significant discriminative ability of VI in predicting death was observed in those children whose VI did not improve by study day 3 (survivors exhibited VI between 30 and 35 throughout this time period, whereas nonsurvivors exhibited VI rising over 57 on day 3) (16, 64). The importance of the severity of VI on day 3 of lung injury was also demonstrated in a study of children with malignancy and ARDS (19). End-tidal alveolar dead space fraction (AVDSf) has also been shown to be independently predictive of mortality within the first day in children with acute hypoxemic respiratory failure even after controlling for severity of illness with Pediatric Risk of Mortality (PRISM) III (OR, 1.83; 95% CI, 1.23–2.7; p = 0.003) (11).
Measurements of Simultaneous Oxygenation and Ventilation Indices.
As our understanding of PARDS pathophysiology improves, it is important to recognize the incremental improvement in outcome prediction by measures that include both oxygenation and ventilation indices. The use of OSI or OI and AVDSf together can improve discrimination with area under the curve greater than 0.7, and the discriminative ability when using P/F or SaO2FIO2 and AVDSf is also quite acceptable (11) (Fig. 4).
Going forward, however, PARDS investigators should examine and report findings at similar time periods through the course of PARDS using all of the above-mentioned indices. Specifically, we recommend studying indices at the onset of PARDS and at 24, 48, and 72 hours after onset of the disease as the utility of these measures in predicting outcome for clinical purposes as well as risk stratification for research studies is most salient at these time points. Furthermore, assessing patients after these time points likely eliminates the “extremes” of the patient populations, that is, those patients that survived and were discharged from the PICU or those that died.
2.2.1 Of the measures available at the bedside, both oxygenation defect and ventilation defect have generally been found to be associated with outcome. There is great inconsistency in the literature, however, concerning the optimal timing of these measurements. We recommend evaluating respiratory indices and biomarkers at the onset of PARDS, within the first 24 hours of onset, as well as serial measures beyond that as indicated according to treatment and/or clinical studies. Strong agreement
2.2.2 For disease severity measures that can be made at the bedside, we recommend that future research studies evaluating both trajectory of illness and recovery should use standardized, minimal datasets with adequately explicit definitions. Strong agreement
2.2.3 Recent adult studies evaluating the effect of dead space ventilation, thereby reflecting lung perfusion, have been highly predictive of outcome. We recommend that future multicenter studies should examine the association of dead space and outcome of PARDS. Strong agreement
Measurements of Ventilator Parameters.
Various ventilator parameters, including peak inspiratory pressures (PIP), positive end-expiratory pressures (PEEP), mean airway pressures (MAP), and tidal volume (TV) have been reported to have predictive value in children with ARDS. As these measures are decidedly clinician dependent and can be considered therapies rather than independent predictors, their utility in multivariate regression analyses is limited. An increased risk of death has been reported in those patients with increased PIP, PEEP, and MAP on univariate analysis (5, 10, 18, 19, 61, 67) and multivariate regression independent of severity of illness (5). Interestingly, Khemani et al (61) studied 398 intubated patients with P/F ratio less than 300, of which 192 had bilateral infiltrates on chest radiograph. TV between 6 and 10 mL/kg was not associated with mortality on multivariate analysis, but higher TVs on day 1 were associated with greater ventilator-free days (OR, 0.88; 95% CI, 0.78–0.98); however, this did not reach statistical significance in the group with bilateral infiltrates (OR, 0.82; 95% CI, 0.70–0.96) (61). Zhu et al (28) also found no association between TVs less than 6 mL/kg to more than 10 mL/kg and mortality or ventilator-free days in a recent study of 345 children with acute hypoxemic respiratory failure. Although seemingly intuitive, most studies do not report associations of ventilation variables and barotrauma; one recent study reported an association between TV and increased risk of barotrauma (7).
2.2.4 Studies examining the relationship between TV, peak airway pressures, PEEP, or MAPs with mortality or length of mechanical ventilation have resulted in conflicting results; some studies exhibit associations with outcomes while others do not. We recommend that future studies incorporating variables such as TV, peak and plateau airway pressures, PEEP, or MAP use explicit protocols and definitions such that these measures can be more robustly evaluated. Strong agreement
Measurements Requiring More In-Depth Calculation
PRISM III and PIM-2.
Other robust measures exist. However, these measures require a much greater collection of patient data and assessment of factors at ICU admission or over the first 12–24 hours of ICU and, therefore, these measures may not necessarily be relevant to the patient in whom PARDS develops 24 hours or more after admission. Specifically, PRISM III and PIM-2 are widely used and validated at predicting outcome, both in children who have ALI/ARDS (2, 9, 10) and in the pediatric critical care population in general. The Pediatric Respiratory Failure Score (15) was proposed to identify the highest risk children with ARDS who might benefit with treatment with extracorporeal membrane oxygenation. This score had a positive predictive value for mortality of 0.72 at onset and 0.7 between 0 and 132 hours (15). These measures essentially assess the extent of multiple organ system dysfunction. Unfortunately, these composite scores include respiratory variables and thereby instill confounding or colinearity in any multivariable analyses intended to dissect out the association of respiratory failure independent of other nonpulmonary organ system failures.
Multiple Organ System Dysfunction.
Of the few studies large enough to allow for multivariate analyses, multiple organ system dysfunction has been the single most important independent clinical risk factor for mortality in children at the onset of ALI/ARDS (5, 9, 10, 17, 68). As expected, progressive increases in the number of failing nonpulmonary organ systems, often termed “new and progressive multiple organ dysfunction,” in the days prior to the onset of ALI/ARDS all indicate progressively increased risk of mortality (Table 1) (5, 7, 10, 17, 20, 67). These findings have been cited to justify the use of noninvasive ventilatory support for ALI/ARDS in children as safest in those children with single organ system failure (68).
Various studies evaluate individual organ system failures. However, comparison is challenging given that there is no consensus to the exact definition of each individual organ system among these studies. For example, renal dysfunction is defined as an elevation in creatinine in some studies compared with the need for renal replacement therapy in other studies; renal dysfunction has been demonstrated to be a strong independent risk factor for mortality in the pediatric oncology population even on multivariable analysis (14, 17, 20). Given the variability in these definitions and the importance of evaluating their association with mortality, we believe there should be great effort in explicitly defining and validating these nonpulmonary organ failures in critically ill children, particularly, in future studies of PARDS.
2.2.5 Among measures requiring more in-depth calculation, we recommend that the use of an estimate of multiple organ system failure should be included in any studies of clinical risk factors associated with outcome in PARDS patients. Strong agreement
2.2.6 While evaluating risk factors related to organ failure in research related to PARDS, caution should be exercised in the use of organ failure scoring systems that include indices of respiratory failure. We recommend the development of a validated, nonpulmonary organ failure definition for use in PARDS research. Strong agreement
Measurements of Biochemical Markers
Many biochemical markers have been measured in critically ill children, though to a lesser degree than those measured in adults, and few studies have focused on children with PARDS. Elevations of von Willebrand factor antigen (68) and endothelin-1, reflecting endothelial injury; elevation of soluble intercellular adhesion molecule-1 (69, 70), reflecting macrophage activation and endothelial cell injury; and elevation in plasminogen activation inhibitor (71), reflecting dysregulated coagulation, resulting in alveolar fibrin deposition (Table 2), have all been associated with increased risk of mortality and prolonged need for mechanical ventilation in children. Elevations of beta natriuretic peptide have also been demonstrated to correlate with worse clinical outcome, but the magnitude of the elevation is to a far lesser degree than that seen in patients with myocardial dysfunction (78). Unlike studies in adult ARDS (72–74) and pediatric sepsis (75, 76), the use of various combinations of biomarkers in providing even stronger prediction of outcome has not been adequately examined in PARDS.
One of the potential values of the study of these biomarkers is in the development of diagnostic testing with greater sensitivity and specificity. If such testing were available as point of care testing, it could help in the risk stratification of patients since these markers are elevated at the onset of PARDS and may better predict outcomes. Elevations of these markers in the earliest days of lung injury also indicate that the pathophysiological process of PARDS is well underway often by the time that the clinician has recognized the extent of disease and certainly before these patients are approached for participation in clinical trials. This latter scenario is relevant given that most therapeutic trials are tested in animal studies within hours after the insult leading to the development of ARDS; by contrast, human subjects enrolling in clinical therapeutic trials are often consented and the investigational therapy is begun often days after the onset of the pathophysiological processes in PARDS. Insomuch as oxygenation defect, dead space ventilation and biomarker changes can be surrogates for the extent of lung tissue damage, the greater these measures deviate from “normal,” the greater the extent of lung injury and, by extension, greater risk of mortality or prolonged mechanical ventilation.
2.2.7 We recommend further research into the potential use of combinations of biomarker levels in providing a stronger prediction outcome. Strong agreement
Assessment of Early Response to Therapy
Early response to therapy has been demonstrated to predict outcome, although the definition of “early” has not been standardized in either clinical practice or research, therefore making comparison of studies difficult. An inability to rapidly reduce FIO2 is a simple, yet sensitive prognostic indicator of poor outcome (13). In addition, a significant association between response to inhaled nitric oxide treatment and outcome (r = 0.43; p < 0.02) has also been demonstrated; all children with a less than 15% improvement in OI did not survive in contrast to 61% of children with a more than 30% improvement in OI surviving (79).
Improvement in oxygenation indices within 24 hours of initiation of high-frequency oscillatory ventilation (HFOV) has also been associated with improved outcome. Conversely, failure of patients to improve after initiation of HFOV has been predictive of mortality. A combination of initial OI greater than 20 and failure to decrease this OI by greater than 20% within 6 hours of start of HFOV predicted death with 86% sensitivity and 83% specificity (80). P(A-a)O2 at 6 hours, oxygenation indices at 24 hours, and MAPs at 36 hours were significantly different in survivors compared with nonsurvivors after initiation of HFOV (81). Persistently high OI after 24 hours of HFOV is also predictive of increased mortality in children receiving HFOV (82).
Certainly, the early management of children with severe lung injury can directly influence their long-term outcome. There is general consensus that lung-protective mechanical ventilation and judicious management of both protein-rich blood product transfusions, especially plasma and platelets, and fluid therapy with the target of attaining euvolemia are associated with improved outcome. Importantly, disregard for these strategies has been associated with worse outcome with a clear association between TV and prevalence of barotrauma (7). These strategies are discussed in further detail later in this PARDS Consensus document (83).
2.2.8 We recommend that early response to therapy should not be used as a primary outcome measure in phase III clinical research trials. Future research should explore the relationship of early response to therapy as an intermediate process variable linked to more clinically relevant, long-term outcomes (e.g., ventilator-free days and mortality). Strong agreement
AREAS OF FUTURE RESEARCH
The pediatric patient “at risk” for PARDS has been sorely understudied, especially when compared to the wealth of literature available in the adult populations. This is important as this group represents a cohort that may benefit from preventative measures that would minimize morbidity and mortality. Of the remaining factors associated with the severity of disease, pediatrics has seen an increase in the volume of research but still not to the same extent as seen in adults. Clearly, a better understanding of the mechanisms of disease in children can result in better management strategies going forward. As the pediatric critical care population is fewer in numbers than the adult critical care population, collaboration and the joining of datasets will be vital to achieve the power necessary to draw robust conclusions and improve the care of critically ill children.
1. Bernard GR, Artigas A, Brigham KL, et al. The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med. 1994;149:818–824
2. Zimmerman JJ, Akhtar SR, Caldwell E, et al. Incidence and outcomes of pediatric acute lung injury. Pediatrics
3. Bindl L, Dresbach K, Lentze MJ. Incidence of acute respiratory distress syndrome
in German children and adolescents: A population-based study. Crit Care Med. 2005;33:209–312
4. Kneyber MC, Brouwers AG, Caris JA, et al. Acute respiratory distress syndrome
: Is it underrecognized in the pediatric intensive care unit? Intensive Care Med. 2008;34:751–754
5. Erickson S, Schibler A, Numa A, et al.Paediatric Study Group; Australian and New Zealand Intensive Care Society. Acute lung injury in pediatric intensive care in Australia and New Zealand: A prospective, multicenter, observational study. Pediatr Crit Care Med. 2007;8:317–323
6. Hu X, Qian S, Xu F, et al.Chinese Collaborative Study Group for Pediatric Respiratory Failure. Incidence, management and mortality of acute hypoxemic respiratory failure and acute respiratory distress syndrome
from a prospective study of Chinese paediatric intensive care network. Acta Paediatr. 2010;99:715–721
7. López-Fernández 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 respiratory distress syndrome
in children. Crit Care Med. 2012;40:3238–3245
8. Sapru S, Flori H, Quasney MW, et al.for the Pediatric Acute Lung Injury Consensus Conference Group. Pathobiology of Acute Respiratory Distress Syndrome
. Pediatr Crit Care Med. 2015;16(Suppl):S6–S22
9. Flori HR, Glidden DV, Rutherford GW, et al. Pediatric acute lunginjury: Prospective evaluation of risk factors associated with mortality. Am J Respir Crit Care Med
10. Dahlem P, van Aalderen WM, Hamaker ME, et al. Incidence and short-termoutcome of acute lung injury in mechanically ventilated children. EurRespir J
11. Ghuman AK, Newth CJ, Khemani RG. The association between theend tidal alveolar dead space fraction and mortality in pediatric acutehypoxemic respiratory failure. Pediatr Crit Care Med
12. Vernon DD, Holzman BH, Lewis P, et al. Respiratory failure in childrenwith acquired immunodeficiency syndrome and acquired immunodeficiencysyndrome-related complex. Pediatrics
13. DeBruin W, Notterman DA, Magid M, et al. Acute hypoxemic respiratoryfailure in infants and children: Clinical and pathologic characteristics. Crit Care Med
14. Nichols DG, Walker LK, Wingard JR, et al. Predictors of acute respiratoryfailure after bone marrow transplantation in children. Crit CareMed
15. Timmons OD, Havens PL, Fackler JC. Predicting death in pediatricpatients with acute respiratory failure. Pediatric Critical Care Study Group.Extracorporeal Life Support Organization. Chest
16. Paret G, Ziv T, Barzilai A, et al. Ventilation index and outcome in childrenwith acute respiratory distress syndrome
17. Keenan HT, Bratton SL, Martin LD, et al. Outcome of children whorequire mechanical ventilatory support after bone marrow transplantation. Crit Care Med
18. Davis SL, Furman DP, Costarino AT Jr. Adult respiratory distress syndromein children: Associated disease, clinical course, and predictorsof death. J Pediatr. 1993;123:35–45
19. Ben-Abraham R, Weinbroum AA, Augerten A, et al. Acute respiratorydistress syndrome in children with malignancy Can we predict outcome? J Crit Care. 2001;16:54–58
20. Lamas A, Otheo E, Ros P, et al. Prognosis of child recipients of hematopoieticstem cell transplantation requiring intensive care. IntensiveCare Med. 2003;29:91–96
21. Willson DF, Thomas NJ, Markovitz BP, et al. Pediatric Acute LungInjury and Sepsis Investigators: Effect of exogenous surfactant (calfactant)in pediatric acute lung injury: A randomized controlled trial. JAMA. 2005;293:470–476
22. Elbahlawan LM, Morrison RR, Jeha S, et al. Impact of neutrophilrecovery on oxygenation in pediatric oncology patients with acutehypoxemic respiratory failure. J PediatrHematolOncol. 2011;33:e296–e299
23. Erickson SE, Shlipak MG, Martin GS, et al. National Institutes ofHealth National Heart, Lung, and Blood Institute Acute RespiratoryDistress Syndrome Network: Racial and ethnic disparities in mortalityfrom acute lung injury. Crit Care Med. 2009;37:1–6
24. Bhadade RR, de Souza RA, Harde MJ, et al. Clinical characteristicsand outcomes of patients with acute lung injury and ARDS. J Postgrad Med. 2011;57:286–290
25. Heffernan DS, Dossett LA, Lightfoot MA, et al. Gender and acuterespiratory distress syndrome in critically injured adults: A prospectivestudy. J Trauma. 2011;71:878–883; discussion 883–885
26. Sigurdsson MI, Sigvaldason K, Gunnarsson TS, et al. Acute respiratorydistress syndrome: Nationwide changes in incidence, treatment andmortality over 23 years. ActaAnaesthesiolScand. 2013;57:37–45
27. Srinivasan R, Asselin J, Gildengorin G, et al. A prospective studyof ventilator-associated pneumonia in children. Pediatrics
28. Zhu YF, Xu F, Lu XL, et al. Chinese Collaborative Study Group forPediatric Hypoxemic Respiratory Failure: Mortality and morbidity ofacute hypoxemic respiratory failure and acute respiratory distresssyndrome in infants and young children. Chin Med J (Engl). 2012;125:2265–2271
29. Adamzik M, Frey U, Sixt S, et al. ACE I/D but not AGT (-6)A/G polymorphismis a risk factor for mortality in ARDS. EurRespir J. 2007;29:482–488
30. Marshall RP, Webb S, Bellingan GJ, et al. Angiotensin convertingenzyme insertion/deletion polymorphism is associated with susceptibilityand outcome in acute respiratory distress syndrome
. Am JRespir Crit Care Med. 2002;166:646–650
31. Jerng JS, Yu CJ, Wang HC, et al. Polymorphism of the angiotensin-convertingenzyme gene affects the outcome of acute respiratory distresssyndrome. Crit Care Med. 2006;34:1001–1006
32. Su L, Zhai R, Sheu CC, et al. Genetic variants in the angiopoietin-2gene are associated with increased risk of ARDS. Intensive CareMed. 2009;35:1024–1030
33. Meyer NJ, Li M, Feng R, et al. ANGPT2 genetic variant is associatedwith trauma-associated acute lung injury and altered plasmaangiopoietin-2 isoform ratio. Am J Respir Crit Care Med. 2011;183:1344–1353
34. Adamzik M, Frey UH, Riemann K, et al. Factor V Leiden mutation isassociated with improved 30-day survival in patients with acute respiratorydistress syndrome. Crit Care Med. 2008;36:1776–1779
35. Lagan AL, Quinlan GJ, Mumby S, et al. Variation in iron homeostasisgenes between patients with ARDS and healthy control subjects. Chest. 2008;133:1302–1311
36. Reiner AP, Wurfel MM, Lange LA, et al. Polymorphisms of the IL1-receptor antagonist gene (IL1RN) are associated with multiple markersof systemic inflammation. ArteriosclerThrombVascBiol. 2008;28:1407–1412
37. Marshall RP, Webb S, Hill MR, et al. Genetic polymorphisms associatedwith susceptibility and outcome in ARDS. Chest. 2002;121:68S–69S
38. Sutherland AM, Walley KR, Manocha S, et al. The association of interleukin6 haplotype clades with mortality in critically ill adults. ArchIntern Med. 2005;165:75–82
39. Nonas SA, Finigan JH, Gao L, et al. Functional genomic insights intoacute lung injury: Role of ventilators and mechanical stress. Proc AmThoracSoc. 2005;2:188–194
40. Flores C, Ma SF, Maresso K, et al. IL6 gene-wide haplotype is associatedwith susceptibility to acute lung injury. Transl Res. 2008;152:11–17
41. Gong MN, Zhou W, Williams PL, et al. Polymorphisms in the mannosebinding lectin-2 gene and acute respiratory distress syndrome
. CritCare Med. 2007;35:48–56
42. Garcia-Laorden MI, Sole-Violan J, Rodriguez de Castro F, et al. Mannose-binding lectin and mannose-binding lectin-associated serineprotease 2 in susceptibility, severity
, and outcome of pneumonia inadults. J Allergy ClinImmunol. 2008;122:368–374, 374.e1–374.e2
43. Gao L, Grant A, Halder I, et al. Novel polymorphisms in the myosinlight chain kinase gene confer risk for acute lung injury. Am J RespirCell MolBiol. 2006;34:487–495
44. Christie JD, Ma SF, Aplenc R, et al. Variation in the myosin light chainkinase gene is associated with development of acute lung injury aftermajor trauma. Crit Care Med. 2008;36:2794–2800
45. Sapru A, Hansen H, Ajayi T, et al. 4G/5G polymorphism of plasminogenactivator inhibitor-1 gene is associated with mortality in intensivecare unit patients with severe pneumonia. Anesthesiology. 2009;110:1086–1091
46. Ye SQ, Simon BA, Maloney JP, et al. Pre-B-cell colony-enhancing factoras a potential novel biomarker in acute lung injury. Am J Respir CritCare Med. 2005;171:361–370
47. Christie JD, Wurfel MM, Feng R, et al. Trauma ALI SNP Consortium(TASC) investigators: Genome wide association identifies PPFIA1 asa candidate gene for acute lung injury risk following major trauma. PLoS One. 2012;7:e28268
48. Wurfel MM, Gordon AC, Holden TD, et al. Toll-like receptor 1 polymorphismsaffect innate immune responses and outcomes in sepsis. Am J Respir Crit Care Med. 2008;178:710–720
49. Gong MN, Zhou W, Williams PL, et al. -308GA and TNFB polymorphismsin acute respiratory distress syndrome
. EurRespir J. 2005;26:382–389
50. Medford AR, Keen LJ, Bidwell JL, et al. Vascular endothelial growthfactor gene polymorphism and acute respiratory distress syndrome
. Thorax. 2005;60:244–248
51. Zhai R, Gong MN, Zhou W, et al. Genotypes and haplotypes of theVEGF gene are associated with higher mortality and lower VEGFplasma levels in patients with ARDS. Thorax. 2007;62:718–722
52. Patwari PP, O’Cain P, Goodman DM, et al. Interleukin-1 receptorantagonist intron 2 variable number of tandem repeats polymorphismand respiratory failure in children with community-acquired pneumonia. Pediatr Crit Care Med. 2008;9:553–559
53. Dahmer MK, O’Cain P, Patwari PP, et al. The influence of geneticvariation in surfactant protein B on severe lung injury in black children. Crit Care Med. 2011;39:1138–1144
54. Baughn JM, Quasney MW, Simpson P, et al. Association of cysticfibrosis transmembrane conductance regulator gene variants withacute lung injury in African American children with pneumonia. CritCare Med. 2012;40:3042–3049
55. Dahlem P, van A, alderen WM, Bos AP. Pediatric acute lung injury. PaediatrRespir Rev. 2007;8:348–362
56. Kendirli T, Demirkol D, Yildizdas D, et al. Critically ill children withpandemic influenza (H1N1) in pediatric intensive care units in Turkey. Pediatr Crit Care Med. 2012;13:e11–e17
57. Zhong NS, Zheng BJ, Li YM, et al. Epidemiology and cause ofsevere acute respiratory syndrome (SARS) in Guangdong, People’sRepublic of China, in February, 2003. Lancet. 2003;362:1353–1358
58. Randolph AG, Vaughn F, Sullivan R, et al. Critically ill children duringthe 2009–2010 influenza pandemic in the United States. Pediatrics
59. Nadel S, Goldstein B, Williams MD, et al. Researching severe sepsisand organ dysfunction in children: A global perspective (RESOLVE)study group. Lancet. 2001;369:836–843
60. Li Y, Wang Q, Chen H, et al. Epidemiological features and risk factoranalysis of children with acute lung injury. World J Pediatr. 2012;8:43–46
61. Khemani RG, Conti D, Alonzo TA, et al. Effect of tidal volume in childrenwith acute hypoxemic respiratory failure. Intensive Care Med. 2009;35:1428–1437
62. Thomas NJ, Shaffer ML, Willson DF, et al. Defining acute lung diseasein children with the oxygenation saturation index. Pediatr Crit CareMed. 2010;11:12–17
63. Khemani RG, Thomas NJ, Venkatachalam V, et al. Pediatric AcuteLung Injury and Sepsis Network Investigators (PALISI): Comparisonof SpO2 to PaO2 based markers of lung disease severity
for childrenwith acute lung injury. Crit Care Med
64. Khemani RG, Smith LS, Zimmerman JJ, et al.for the Pediatric Acute Lung Injury Consensus Conference Group. Pediatric Acute Respiratory Distress Syndrome
: Definition, Incidence, and Epidemiology: Proceedings From the Pediatric Acute Lung Injury Consensus Conference. Pediatr Crit Care Med. 2015;16(Suppl):S23–S40
65. Paret G, Ziv T, Augarten A, et al. Acute respiratory distress syndromein children: A 10 year experience. Isr Med Assoc J. 1999;1:149–153
66. Nuckton TJ, Alonso JA, Kallet RH, et al. Pulmonary dead-space fractionas a risk factor for death in the acute respiratory distress syndrome
. N Engl J Med. 2002;346:1281–1286
67. Peters MJ, Tasker RC, Kiff KM, et al. Acute hypoxemic respiratory failurein children: Case mix and the utility of respiratory severity
indices. Intensive Care Med. 1998;24:699–705
68. Piastra M, De L, uca D, Marzano L, et al. The number of failing organspredicts non-invasive ventilation failure in children with ALI/ARDS. Intensive Care Med. 2011;37:1510–1516
69. Flori HR, Ware LB, Milet M, et al. Early elevation of plasma vonWillebrand factor antigen in pediatric acute lung injury is associatedwith an increased risk of death and prolonged mechanical ventilation. Pediatr Crit Care Med. 2007;8:96–101
70. Samransamruajkit R, Prapphal N, Deelodegenavong J, et al. Plasmasoluble intercellular adhesion molecule-1 (sICAM-1) in pediatricARDS during high frequency oscillatory ventilation: A predictor ofmortality. Asian Pac J Allergy Immunol. 2005;23:181–188
71. Sapru A, Curley MA, Brady S, et al. Elevated PAI-1 is associated withpoor clinical outcomes in pediatric patients with acute lung injury. Intensive Care Med. 2010;36:157–163
72. Ware LB, Koyama T, Billheimer DD, et al. NHLBI ARDS ClinicalTrials Network: Prognostic and pathogenetic value of combining clinicaland biochemical indices in patients with acute lung injury. Chest. 2010;137:288–296
73. Ong T, McClintock DE, Kallet RH, et al. Ratio of angiopoietin-2 toangiopoietin-1 as a predictor of mortality in acute lung injury patients. Crit Care Med. 2010;38:1845–1851
74. Dolinay T, Kim YS, Howrylak J, et al. Inflammasome-regulated cytokinesare critical mediators of acute lung injury. Am J Respir Crit CareMed. 2012;185:1225–1234
75. Briassoulis G, Venkataraman S, Thompson A. Cytokines and metabolicpatterns in pediatric patients with critical illness. Clin DevImmunol. 2010;2010:354047
76. Wong HR, Salisbury S, Xiao Q, et al. The pediatric sepsis biomarkerrisk model. Crit Care. 2012;16:R174
77. Dobyns EL, Eells PL, Griebel JL, et al. Elevated plasma endothelin-1and cytokine levels in children with severe acute respiratory distresssyndrome. J Pediatr. 1999;135:246–249
78. Reel B, Oishi PE, Hsu JH, et al. Early elevations in B-type natriureticpeptide levels are associated with poor clinical outcomes in pediatricacute lung injury. PediatrPulmonol. 2009;44:1118–1124
79. Goldman AP, Tasker RC, Hosiasson S, et al. Early response toinhaled nitric oxide and its relationship to outcome in childrenwith severe hypoxemic respiratory failure. Chest. 1997;112:752–758
80. Sarnaik AP, Meert KL, Pappas MD, et al. Predicting outcome in childrenwith severe acute respiratory failure treated with high-frequencyventilation. Crit Care Med. 1996;24:1396–1402
81. Lochindarat S, Srisan P, Jatanachai P. Factors effecting the outcomeof acute respiratory distress syndrome
in pediatric patients treatedwith high frequency oscillatory ventilation. J Med Assoc Thai. 2003;86(Suppl 3):S618–S627
82. Arnold JH, Anas NG, Luckett P, et al. High-frequency oscillatory ventilationin pediatric respiratory failure: A multicenter experience. CritCare Med. 2000;28:3913–3919
83. Rimensberger PC, Cheifetz IMfor the Pediatric Acute Lung Injury Consensus Conference Group. . Ventilatory Support in Children With Pediatric Acute Respiratory Distress Syndrome
: Proceedings From the Pediatric Acute Lung Injury Consensus Conference. Pediatr Crit Care Med. 2015;16(Suppl):S51–S60
APPENDIX 1: Pediatric Acute Lung Injury Consensus Conference Group
Organizing Committee: Philippe Jouvet, University of Montreal, Canada; Neal J. Thomas, Pennsylvania State University; Douglas F. Willson, Medical College of Virginia
Section 1, Definition, incidence, and epidemiology: Simon Erickson, Princess Margaret Hospital for Children, Australia; Robinder Khemani, University of Southern California; Lincoln Smith, University of Washington; Jerry Zimmerman, University of Washington
Section 2, Pathophysiology, comorbidities, and severity: Mary Dahmer, University of Michigan; Heidi Flori, Children’s Hospital & Research Center Oakland; Michael Quasney, University of Michigan; Anil Sapru, University of California San Francisco
Section 3, Ventilatory support: Ira M. Cheifetz, Duke University; Peter C. Rimensberger, University Hospital of Geneva, Switzerland
Section 4, Pulmonary-specific ancillary treatment: Martin Kneyber, University Medical Center Groningen, The Netherlands; Robert F. Tamburro, Pennsylvania State University
Section 5, Nonpulmonary treatment: Martha A. Q. Curley, University of Pennsylvania; Vinay Nadkarni, University of Pennsylvania; Stacey Valentine, Harvard University
Section 6, Monitoring: Guillaume Emeriaud, University of Montreal, Canada; Christopher Newth, University of Southern California
Section 7, Noninvasive support and ventilation: Christopher L. Carroll, University of Connecticut; Sandrine Essouri, Université Pierre et Marie Curie, France
Section 8, Extracorporeal support: Heidi Dalton, University of Arizona; Duncan Macrae, Royal Brompton Hospital, United Kingdom
Section 9, Morbidity and long-term outcomes: Yolanda Lopez, Cruces University Hospital, Spain; Michael Quasney, University of Michigan; Miriam Santschi, Université de Sherbrooke, Canada; R. Scott Watson, University of Pittsburgh
Literature Search Methodology: Melania Bembea, Johns Hopkins University