Circulating nucleosomes are associated with mortality in pediatric acute respiratory distress syndrome
Abstract
Mechanisms underlying pediatric acute respiratory distress syndrome (PARDS) are poorly understood. The recent implication of circulating nucleosomes as pathogenic in sepsis and trauma-associated ARDS in adults led us to investigate the significance of nucleosomes in PARDS. We conducted a prospective, observational study on children with PARDS at the Children's Hospital of Philadelphia between July 2014 and September 2015. Plasma was collected within 48 h of PARDS onset and nucleosomes quantified by enzyme-linked immunosorbent assay. Samples from 76 children with PARDS (11 deaths, 14%) were collected early [median 15 (IQR 7, 21) h] after PARDS onset. Nucleosome levels were higher in nonsurvivors [0.59 AU (IQR 0.46, 0.84)] relative to survivors [0.21 AU (IQR 0.08, 0.33), rank sum P < 0.001]. Nucleosome levels were not associated with either Berlin (P = 0.845) or PALICC (P = 0.886) oxygenation categories, nor with etiology of PARDS (P = 0.527). Nucleosomes were correlated with increasing numbers of nonpulmonary organ failures (P = 0.009 for trend), and were higher in patients whose PaO2/FiO2 worsened (P = 0.012) over the first 72 h of PARDS. In regression analysis, nucleosome levels were independently associated with mortality after adjusting for either age, severity of illness score, number of nonpulmonary organ failures, vasopressor score, or PaO2/FiO2 (all P < 0.05). In conclusion, plasma nucleosome levels in early PARDS were associated with increased mortality, correlated with number of nonpulmonary organ failures, and preceded worsening oxygenation. The potential utility of this biomarker for prognostication, risk stratification, and mechanistic insight should be investigated further.
mechanisms underlying pediatricacute respiratory distress syndrome (PARDS) remain elusive. While PARDS was historically defined by adult ARDS criteria (6, 25), it possesses a distinct epidemiologic, comorbidity, and outcome profile, prompting the Pediatric Acute Lung Injury Consensus Conference (PALICC) to develop pediatric-specific definitions in 2015 (23). Among other differences, the PALICC definition of PARDS uses oxygenation index (OI), rather than PaO2/FiO2, for risk stratification, despite the inconsistent relationship between oxygenation and outcome (1, 31). Common to both adult (35, 36) and pediatric (13, 39) ARDS is the significance of nonpulmonary organ failures as a predictor of poor outcome. While several investigations have focused on the systemic inflammatory response and the associated neutrophil activation and cytokine release (9, 36), the pathogenic mechanisms invoked remain incomplete for explaining the development of lung injury and multisystem organ failure (MSOF).
Recent studies in adult ARDS have implicated circulating nucleosomes, the histone/DNA complexes resulting from nuclear chromatin degradation released after cellular damage, as potentially pathogenic in sepsis (4, 12, 24, 38), aspiration (41), and trauma-related ARDS (2). Normally located within the nucleus, nucleosomes released into the circulation act as damage-associated molecular patterns (DAMP), and have been shown to be toxic to multiple cell types (2, 12, 24), offering a novel mechanism linking diverse inciting insults with subsequent lung injury and organ failure. However, whether nucleosomes are associated with PARDS development or progression is unknown.
Rapid identification of children with PARDS who are most at risk of a poor outcome is essential for accurate risk stratification in clinical trials, as it allows redirection of aggressive interventions toward the sickest cohort. Unlike adult ARDS, few studies address the utility of biomarkers in PARDS. The reduced incidence of PARDS [12.8 per 100,000 person-years; (42)] relative to adults [78.9 per 100,000 person-years; (28)] and the lower mortality in pediatrics (22, 39, 42), complicates the design of studies associating biomarkers to clinically relevant outcomes.
Given recent data implicating circulating nucleosomes in adult ARDS, we sought to determine their relevance in a cohort of children with PARDS. We hypothesized that nucleosome levels would be elevated in in PARDS nonsurvivors relative to survivors.
METHODS
Study design and patient selection.
This prospective, observational study was approved by the Children's Hospital of Philadelphia's (CHOP) Institutional Review Board, and written informed consent was obtained from caregivers prior to enrollment. Clinical data were collected prospectively.
Consecutive patients in the pediatric intensive care unit (PICU) were screened daily for Berlin-defined ARDS and eligibility between July 1, 2014, and September 30, 2015. Inclusion criteria were 1) acute respiratory failure requiring invasive (via endotracheal tube) mechanical ventilation projected to last >24 h, 2) invasive arterial access, 3) age > 1 mo (to avoid confounding by neonatal physiology) and < 18 years, 4) PaO2/FiO2 ≤ 300 on two consecutive arterial blood gases separated by ≥1 h on positive end-expiratory pressure (PEEP) ≥ 5 cmH2O, and 5) bilateral parenchymal infiltrates on radiograph. Exclusion criteria were 1) respiratory failure primarily from cardiac failure (determined by echocardiography), 2) exacerbation of underlying chronic respiratory disease, 3) chronic ventilator dependence, 4) mixing cyanotic heart disease, 5) mechanical ventilation for >7 days before PaO2/FiO2 ≤ 300, 6) ARDS established outside of the CHOP PICU, and 7) inability to obtain informed consent.
Determination of bilateral infiltrates was made independently by a blinded PICU attending and a blinded pediatric radiologist; only cases agreed upon as consistent with Berlin ARDS criteria met inclusion. As the study was initiated prior to the 2015 PALICC definitions of PARDS, we did not screen patients based on OI; however, all patients met PARDS criteria in addition to Berlin ARDS criteria.
At-risk intubated controls.
To facilitate comparisons with intubated patients without PARDS, an additional convenience sample of mechanically ventilated children with risk factors for PARDS who were screened for the study, but who did not meet Berlin (or PALICC) oxygenation criteria (PaO2/FiO2 > 300), were used as intubated at-risk controls.
Plasma collection and measurements.
Blood was collected within 48 h of PARDS onset (defined as time of meeting all Berlin criteria) in citrated tubes (Becton, Dickinson and Company, Franklin Lakes, NJ), centrifuged within 30 min of collection (2,000 g, 20 min, 20°C) to generate platelet-poor plasma, aliquoted to prevent freeze/thaw cycles, and stored at −80°C. Nucleosomes were measured in duplicate using an enzyme-linked immunosorbent assay (Cell Death detection ELISAPLUS, Roche, Basel, Switzerland). Replicates all had variation < 5%. As a standard curve was not provided, nucleosome concentrations are reported as arbitrary units (AU), which are values obtained for each sample normalized to the positive control included in the kit. Interassay coefficient of variation was 5.2%, with a lower limit of detection of 0.01 AU.
Equations and definitions.
Metrics of oxygenation utilized were PaO2/FiO2 and OI {[mean airway pressure (mPaw) × FiO2 × 100]/PaO2} at PARDS onset, and 24 and 72 h later. The vasopressor score (15, 37) is dopamine dose (μg·kg−1·min−1) × 1 + dobutamine (μg·kg−1·min−1) × 1 + epinephrine (μg·kg−1·min−1) × 100 + norepinephrine (μg·kg−1·min−1) × 100 + phenylephrine (μg·kg−1·min−1) × 100 + milrinone (μg·kg−1·min−1) × 10 + vasopressin (U·kg−1·min−1) × 10,000. Nonpulmonary organ failures at time of ARDS diagnosis were identified by accepted definitions in children (16). Severity of illness score used is the Pediatric Risk of Mortality (PRISM) III at 12 h.
The primary outcome reported was PICU mortality, with the “cause of death” determined by the treating physician. We also reported ventilator-free days (VFD) at 28 days and length of mechanical ventilation. All mention of “mechanical ventilation” in this study implied “invasive” ventilation, and noninvasive support was not counted toward VFD or total ventilator days. For VFD and duration of mechanical ventilation, the first day was initiation of invasive ventilation. Liberation from invasive ventilation for >24 h defined duration of mechanical ventilation. Patients requiring reinitiation of invasive ventilation after 24 h of extubation had the extra days counted toward total ventilator days. VFD were determined by subtracting total ventilator days from 28 in survivors. All patients with total ventilator days ≥ 28 days, and all PICU nonsurvivors were assigned VFD = 0.
Statistical analysis.
The majority of data were nonnormally distributed as measured by Shapiro-Wilks, and are reported as median [interquartile range (IQR)], and differences between groups compared using nonparametric statistics. Cuzick's nonparametric test of trend was used to assess for monotonic trends across ordered groups (10). Categorical data were compared by Fisher exact test. To assess predictive ability of nucleosomes for mortality, the area under the receiver operating characteristic (AUROC) curve was computed. AUROC for different predictors were compared based on the methods of DeLong et al. (11). To test association of nucleosomes with mortality, bivariate logistic regression was performed retaining nucleosome levels, with potential confounders included one at a time in the model. The confounders tested (age, PRISM III, organ failures, vasopressor score, and PaO2/FiO2) were chosen for univariate association with mortality at P ≤ 0.2. Malignancy was tested given the known association of malignancy with elevated nucleosome levels. As this was a pilot study, we did not perform a sample size estimation a priori. Analysis was performed with Stata/SE 14 (College Station, TX).
RESULTS
Description of the cohort.
During the study period, 327 patients were screened, 102 children met criteria, and 76 cases with PARDS were enrolled (Fig. 1). The most common etiologies for PARDS were pneumonia (47%) and nonpulmonary sepsis (29%). Blood was drawn at a median 15 (IQR 7, 21) h after PARDS onset; 67 of 76 (88%) samples were collected ≤24 h after PARDS onset. Nucleosome levels were not correlated with either patient age (Spearman ρ 0.01, P = 0.922) or time of blood draw relative to PARDS onset (ρ = −0.04, P = 0.705).
Fig. 1.Flow diagram of study screening and eligibility.
Association of nucleosomes with mortality.
Of the 76 PARDS cases, there were 11 nonsurvivors (14%). Nonsurvivors died at a median of 6 (IQR 4, 16; range 2 to 35) days after plasma collection, and no patients died within 48 h of plasma collection. Nonsurvivors had worse PRISM III scores and more nonpulmonary organ failures, but similar measures of lung injury at PARDS onset, compared with survivors (Table 1). Nucleosome levels were higher in nonsurvivors [0.59 AU (IQR 0.46, 0.84)] relative to survivors [0.21 AU (IQR 0.08, 0.33), rank sum P < 0.001, Fig. 2A]. No association was seen between nucleosome levels and cause of death (ANOVA on ranks P = 0.866, Fig. 2B).
| Variable | Survivors (65)† | Nonsurvivors (11)† | P Value |
|---|---|---|---|
| Age, yr | 4.3 [1.5, 12.6] | 1.3 [0.5, 8.6] | 0.165 |
| Female/male, %/% | 31/34 (48/52) | 6/5 (55/45) | 0.518 |
| PRISM III at 12 h | 10 [6, 16] | 17 [11, 33] | 0.025 |
| Malignancy | 10 (15) | 2 (18) | 0.553 |
| Cause of PARDS | 0.450 | ||
| Aspiration pneumonia | 8 (12) | 1 (9) | |
| Infectious pneumonia | 32 (49) | 4 (36) | |
| Nonpulmonary sepsis | 18 (28) | 4 (36) | |
| Trauma | 4 (6) | 0 | |
| Other | 3 (5) | 2 (18) | |
| Nonpulmonary organ failures at PARDS onset | 1 [1, 3] | 4 [2, 5] | <0.001 |
| Vasopressor score | 9 [2, 15] | 15 [5, 32] | 0.112 |
| At PARDS onset | |||
| PaO2/FiO2 | 165 [103, 235] | 225 [130, 262] | 0.144 |
| OI | 10.4 [6, 17.9] | 7.7 [7.1, 16.9] | 0.965 |
| PEEP, cmH2O | 10 [7, 12] | 12 [9, 13] | 0.233 |
| Peak pressure, cmH2O | 31 [26, 35] | 31 [27, 34] | 0.971 |
| Tidal volume, ml/kg | 7.6 [6.3, 8.8] | 7.5 [6.3, 8] | 0.670 |
| Ancillary therapies | |||
| Inhaled nitric oxide | 22 (34) | 5 (45) | 0.506 |
| Prone positioning | 3 (5) | 0 | 1 |
| Neuromuscular blockade | 29 (45) | 7 (64) | 0.332 |
| Alternative ventilator modes | 20 (31) | 5 (45) | 0.489 |
| ECMO | 5 (8) | 0 | 1 |
Fig. 2.Plasma nucleosome levels in patients with PARDS (n = 76). A: levels between survivors (n = 65) and nonsurvivors (n = 11); P value represents the result of a rank sum test. B: plasma nucleosomes in nonsurvivors stratified by cause of death; multisystem organ failure (MSOF; n = 4), refractory hypoxemia (n = 5), or poor neurologic prognosis (n = 2). P value represents the results of a Kruskal-Wallis ANOVA on ranks.
Confounders were included one at a time in bivariate logistic regression. Nucleosomes remained independently associated with mortality after adjustment for either age, PRISM III, presence of malignancy, number of nonpulmonary organ failures, vasopressor score, or PaO2/FiO2 (Table 2). In a sensitivity analysis restricted to patients without malignancy (n = 64, 9 deaths), nucleosomes retained association with mortality (odds ratio 1.73, 95% CI 1.24 to 2.40, P = 0.001). Nucleosomes demonstrated good discriminative ability for mortality, with an AUROC of 0.82 (95% CI 0.66 to 0.97), outperforming both Berlin and PALICC oxygenation categories (P < 0.01 when comparing AUROC, Table 3).
| Variable* | OR (95% CI) | P Value |
|---|---|---|
| Nucleosomes | 1.52 (1.19 to 1.94) | 0.001 |
| Nucleosomes + age, yr | 1.52 (1.19 to 1.95) | 0.001 |
| Nucleosomes + PRISM III | 1.42 (1.10 to 1.84) | 0.008 |
| Nucleosomes + malignancy | 1.59 (1.21 to 2.10) | 0.001 |
| Nucleosomes + nonpulmonary organ failures | 1.36 (1.05 to 1.75) | 0.041 |
| Nucleosomes + vasopressor score | 1.49 (1.16 to 1.92) | 0.002 |
| Nucleosomes + PaO2/FiO2 | 1.52 (1.18 to 1.95) | 0.001 |
| Variable | AUROC (95% CI) | P Value |
|---|---|---|
| Berlin (PaO2/FiO2) categories | 0.41 (0.23 to 0.59) | 1 |
| PALICC (OI) categories | 0.44 (0.25 to 0.62) | 1 |
| PRISM III at 12 h* | 0.71 (0.54 to 0.89) | 0.024 |
| Nonpulmonary organ failures* | 0.80 (0.70 to 0.89) | 0.002 |
| Nucleosomes* | 0.82 (0.66 to 0.97) | <0.001 |
Association with organ failures and lung injury.
Nucleosome levels demonstrated no association between oxygenation severity categories at PARDS onset defined by either Berlin (Fig. 3A) or PALICC (Fig. 3B) criteria. There was no association with etiology of PARDS (Fig. 3C), nor when comparing infectious (pneumonia and sepsis) and noninfectious (aspiration, trauma, and others) etiologies (rank sum P = 0.678). There was a strong correlation between nucleosome levels and increasing number of nonpulmonary organ failures (P = 0.009 for trend, Fig. 3D). There was no association between nucleosome levels and the initial peak inflating pressure (PIP) (Spearman ρ = 0.190, P = 0.121), PEEP (ρ = 0.124, P = 0.316), ΔP (PIP minus PEEP, ρ = 0.155, P = 0.208), or tidal volume (ρ = 0.067, P = 0.590).
Fig. 3.Relationship between plasma nucleosome levels and Berlin (A) or PALICC (B) oxygenation categories, PARDS etiologic factor (C), and number of nonpulmonary organ failures (D). P values represent results of either a Kruskal-Wallis or of a nonparametric test of trend.
Patients who had lower PaO2/FiO2 72 h after PARDS onset (relative to initial PaO2/FiO2) had higher nucleosome levels, demonstrating an association between worsening oxygenation over time and nucleosome levels (Fig. 4A). A similar trend was seen for patients with worse OI at 72 h (Fig. 4B). Elevated nucleosomes had modest predictive ability for worsened PaO2/FiO2 at 72 h (Fig. 4C), with a similar trend seen for worsened OI at 72 h (Fig. 4D). The association between nucleosome levels and worse oxygenation persisted after adjustment for potential confounders (Table 4).
Fig. 4.Plasma nucleosome levels stratified by whether PaO2/FiO2 (A) or OI (B) at 72 h after PARDS onset was improved or worsened relative to corresponding values at PARDS onset. P values represent the results of rank sum tests. AUROC curves were constructed to assess the ability of nucleosomes to predict worsening of PaO2/FiO2 (C) or OI (D) at 72 h relative to values at PARDS onset.
| Variable* | OR (95% CI) | P Value |
|---|---|---|
| Worse PaO2/FiO2 at 72 h | ||
| Nucleosomes | 1.25 (1.03 to 1.52) | 0.024 |
| Nucleosomes + age, yr | 1.25 (1.03 to 1.52) | 0.025 |
| Nucleosomes + PRISM III | 1.51 (1.15 to 1.98) | 0.003 |
| Nucleosomes + malignancy | 1.24 (1.01 to 1.51) | 0.039 |
| Nucleosomes + nonpulmonary organ failures | 1.39 (1.08 to 1.78) | 0.011 |
| Nucleosomes + vasopressor score | 1.38 (1.08 to 1.74) | 0.008 |
| Nucleosomes + initial PaO2/FiO2 | 1.27 (1.03 to 1.57) | 0.024 |
| Worse OI at 72 h | ||
| Nucleosomes | 1.20 (0.99 to 1.45) | 0.066 |
| Nucleosomes + age, yr | 1.20 (0.99 to 1.45) | 0.070 |
| Nucleosomes + PRISM III | 1.41 (1.09 to 1.82) | 0.009 |
| Nucleosomes + malignancy | 1.18 (0.97 to 1.44) | 0.104 |
| Nucleosomes + nonpulmonary organ failures | 1.34 (1.05 to 1.72) | 0.020 |
| Nucleosomes + vasopressor score | 1.32 (1.05 to 1.66) | 0.019 |
| Nucleosomes + initial OI | 1.26 (1.02 to 1.55) | 0.034 |
Comparison with at-risk intubated controls.
An additional 23 intubated control patients at risk for PARDS, but not meeting oxygenation criteria, were also enrolled. As expected, PARDS cases had significantly worse nonpulmonary organ failures, vasopressor use, lung injury, and outcomes relative to controls (Table 5). Plasma nucleosome levels were higher in PARDS patients [median 0.25 AU (IQR 0.08, 0.44)] compared with intubated at-risk controls [0.09 AU (IQR 0.06, 0.15), rank sum P < 0.001, Fig. 5]. Confounders with a univariate association with a PARDS diagnosis at P ≤ 0.2 (Table 5) were included one at a time in bivariate logistic regression. The association between nucleosome levels and PARDS diagnosis persisted after adjustment for confounders (Table 6).
| Variable | Control (23)† | PARDS (76)† | P Value* |
|---|---|---|---|
| Age, yr | 8.3 [2.3, 14.1] | 4.1 [1.2, 11.2] | 0.215 |
| Female/male, %/% | 8/15 (35/65) | 37/39 (49/51) | 0.350 |
| PRISM III at 12 h | 9 [5, 12] | 11 [6, 17] | 0.065 |
| Malignancy, % | 7 (30) | 12 (16) | 0.101 |
| Risk factor for intubation, % | <0.001 | ||
| Aspiration pneumonia | 0 | 9 (12) | |
| Infectious pneumonia | 3 (13) | 36 (47) | |
| Nonpulmonary sepsis | 10 (43) | 22 (29) | |
| Trauma | 6 (26) | 4 (5) | |
| Other | 4 (17) | 5 (7) | |
| Nonpulmonary organ failures | 1 [0, 2] | 2 [1, 3] | 0.037 |
| Vasopressor score | 0 [0, 15] | 10 [3, 16] | 0.048 |
| Initial settings | |||
| PaO2/FiO2 | 441 [365, 499] | 178 [110, 243] | <0.001 |
| OI | 2.5 [1.6, 3.7] | 9.6 [6.4, 17.4] | <0.001 |
| PEEP, cmH2O | 6 [5, 10] | 10 [8, 12] | <0.001 |
| Peak pressure, cmH2O | 21 [15, 27] | 31 [26, 35] | <0.001 |
| Tidal volume, ml/kg | 7.5 [6.9, 8.3] | 7.6 [6.3, 8.5] | 0.925 |
| Outcomes | |||
| Ventilator days, all | 4 [1, 7] | 10 [6, 16] | <0.001 |
| Ventilator days, survivors | 4 [1, 7] | 11 [7, 16] | <0.001 |
| VFD at 28 days | 24 [22, 27] | 16 [1, 21] | <0.001 |
| PICU mortality, % | 0 | 11 (14) | 0.040 |
Fig. 5.Plasma nucleosome levels in PARDS cases (n = 76) and intubated, at-risk controls without PARDS (n = 23). P value represents the result of a rank sum test.
| Variable* | OR (95% CI) | P Value |
|---|---|---|
| Nucleosomes | 2.06 (1.28 to 3.31) | 0.003 |
| Nucleosomes + PRISM III | 2.13 (1.28 to 3.54) | 0.003 |
| Nucleosomes + malignancy | 1.98 (1.24 to 3.15) | 0.004 |
| Nucleosomes + diagnosis | 1.91 (1.16 to 3.13) | 0.010 |
| Nucleosomes + nonpulmonary organ failures | 2.04 (1.24 to 3.35) | 0.005 |
| Nucleosomes + vasopressor score | 2.09 (1.27 to 3.41) | 0.003 |
DISCUSSION
This is the first study examining the clinical utility of plasma nucleosomes in PARDS. Plasma nucleosomes were higher in nonsurvivors, correlated with nonpulmonary organ failures, and were associated with subsequent worsening oxygenation. Nucleosomes demonstrated modest predictive ability for PICU mortality and remained independently associated with nonsurvival after adjustment for potential confounders.
We demonstrated a strong correlation between nucleosomes and increasing number of organ failures, which may mediate the association with nonsurvival. In the setting of cell death, nuclear proteins, including the DNA and histone components of chromatin, can be released into circulation. Chromatin is degraded to mono- and oligonucleosomes, and subsequently down to component histone and free DNA (18), where they function as DAMPs. Initial studies focused on the role of nucleosomes in autoimmune disorders (21) and oncologic processes (17). More recently, circulating histones have been implicated as mediators of endothelial and organ dysfunction in sepsis (38), and several clinical studies have demonstrated their relevance to critical illness. Histones and their parent nucleosomes have been implicated in cardiac dysfunction (4, 5), renal injury (12), and dendritic cell necrosis (24). In adults with trauma, nucleosome levels correlated with trauma severity scores, with histone levels correlating with Sequential Organ Failure Assessment scores and subsequent lung injury (2), consistent with our findings in this PARDS cohort.
To our knowledge, this is the first study demonstrating an association between nucleosome levels and mortality in a heterogeneous PARDS cohort. In five studies (3 adult sepsis, 1 adult aspiration ARDS, and 1 pediatric sepsis), nucleosome or histone levels have been demonstrated to be higher in nonsurvivors (4, 12, 24, 40, 41). In a cohort of septic adults (n = 19), elevated histone H4 on day 1 predicted increased mortality at days 28 and 90 (12). Two separate studies demonstrated elevated total circulating histones [n = 65, (4)] and elevated nucleosomes [n = 49, (24)] in nonsurvivors of adult septic shock. In a fourth study, adults with aspiration ARDS (n = 21) had elevated plasma nucleosomes and H4 above control patients, with higher levels seen in nonsurvivors compared with survivors (41). The single pediatric study (40) described a cohort of meningococcal sepsis (n = 35) in which all nonsurvivors died within 48 h of presentation, with nucleosome levels elevated at multiple time points in nonsurvivors relative to survivors.
Few studies exist examining the clinical utility of circulating biomarkers in PARDS, in contrast with the more extensive investigations in adult ARDS (reviewed in Ref. 32). No studies have investigated the role of nucleosomes in PARDS. The lower mortality rate of PARDS (relative to adults), which appears to be decreasing over time (39, 42), makes the selection of clinically relevant outcomes challenging for biomarker studies. Flori et al. have demonstrated the prognostic utility of brain-type natriuretic peptide (27), plasminogen activator inhibitor-1 (30), and soluble intracellular adhesion molecule-1 [sICAM-1, (14)]. Others have validated the association between increased sICAM-1 and nonsurvival in PARDS (3, 29). Elevations in the inflammatory markers C-reactive protein (8) and matrix metalloproteins (20) have also been associated with poor outcomes in PARDS. A single small study suggested utility of the lung epithelial derived Krebs von den Lungen-6 in discriminating nonsurvival in PARDS (7). Our study introduces a new class of circulating biomarker, a DAMP, which has potential therapeutic utility in addition to prognostic value. Nucleosomes and their constituent histones are cytotoxic across multiple cell and organ types. A hypothetically central role for nucleosomes in mediating organ dysfunction (including lung injury) after a variety of disparate types of inciting insults make it an attractive target for inhibition. Prior therapies known to ameliorate histone toxicity, such as heparin (41) and recombinant human activated protein C (38), which do not demonstrate benefit in unselected sepsis (19, 26), may find renewed interest in a targeted, sicker population defined by excess nucleosomes. In pediatrics, the lower overall mortality rate makes survival endpoints for clinical trials impractical. However, stratifying by a validated biomarker may enrich for a sicker cohort with higher mortality. In our cohort, elevated nucleosomes at a median of 15 h after PARDS onset predicted worsening PaO2/FiO2 at 72 h, suggesting that a study stratifying by elevated nucleosome levels could potentially enrich for a persistent PARDS phenotype with a higher mortality. Notably, nucleosome levels did not differ by cause of death. One potential explanation is that most of these patients had severe hypoxemia and MSOF at time of phlebotomy, most of which had recovered at the time of withdrawal of care, other than the brain dysfunction. However, given the small number of patients who died, this remains an intriguing observation in need of further study.
The relationship to mortality in our cohort may be mediated via MSOF. While this current study cannot establish causality between nucleosomes and mortality, existing translational literature suggests a plausible causal relationship. Furthermore, nucleosome levels remain associated with mortality after adjustment for number of organ failures, suggesting additional risk imparted by the presence of nucleosomes. However, the reduced odds ratio after adjustment for organ failure suggests some measure of confounding of the relationship between nucleosomes and mortality, an issue which may be best clarified by animal models of lung injury. Nonpulmonary organ failure has been consistently demonstrated to be a strong risk factor for nonsurvival in other PARDS cohorts (13, 39). Additionally, several animal and cell studies have demonstrated endothelial (2, 12, 38) and organ dysfunction with exogenous histone treatment, including cardiac toxicities (4) and renal failure (38). The mechanism of endothelial and organ injury remains incompletely defined, but in multiple cell types, histone-mediated calcium influx led to increased cell death (2, 5), which may explain the correlation between nucleosomes and organ failure in our cohort.
We did not find an association with pulmonary-specific metrics of injury, such as Berlin or PALICC oxygenation categories at PARDS onset. This is in contrast with a single study of adults with aspiration ARDS, which showed a stepwise increase in plasma nucleosomes across worsening Berlin severity categories (41). It is possible that the elevated nucleosome levels do not mediate their effects through worsening lung injury, or that severity of lung injury (as measured by initial PaO2/FiO2 or OI) is not reflected by increased nucleosome levels. We are reluctant to conclude this from this study alone, especially given the heterogeneity of PARDS etiologies and the prior laboratory data suggesting direct pulmonary toxicity with exogenous histone treatment (2, 38). Furthermore, in patients with worsening PaO2/FiO2 72 h after PARDS onset, nucleosomes were elevated relative to those with improving PaO2/FiO2, suggesting an association between increased nucleosome levels and subsequent worsening lung injury. We also confirm, as in prior cohorts, that initial oxygenation is a poor predictor of outcome in PARDS (22, 33, 34, 39).
Our study has limitations. This was conducted at a single center, and while severity of illness and etiologies of PARDS are similar to other cohorts, these findings may not generalize, and independent validation is a necessity. The mortality rate is low, although comparable to the 13% mortality seen in our recently published cohort defined by similar eligibility criteria (39), which limits the ability to simultaneously adjust for multiple potential confounders. Heterogeneous etiologies of PARDS were included, and future studies could benefit from more homogenous populations to study the prognostic utility within subphenotypes of PARDS. The at-risk intubated control population was significantly less ill than the PARDS cohort, precluding the ability to directly compare these groups. However, after adjustment for confounders, nucleosomes remained independently associated with a diagnosis of PARDS, suggesting potential additional utility of nucleosomes in differentiating PARDS from similarly ill conditions. Bronchoalveolar lavage was not performed, and the presence of nucleosomes in the alveolar compartment could not be assessed. Finally, plasma collection was allowed up to 48 h, although 88% samples were collected ≤24 h after PARDS onset, similar to other studies. It is possible that an earlier timeframe for blood collection would yield more reproducible results; however, several variables, such as PICU admission and PARDS onset, are inherently arbitrary, and more acute collection time may not adequately address this variability. Despite these limitations we demonstrated the independent association of circulating nucleosomes with mortality in PARDS.
In conclusion, plasma nucleosomes were associated with increased mortality in PARDS, correlated with the number of nonpulmonary organ failures, and preceded subsequent worsening of oxygenation. The potential utility of this biomarker for prognostication, risk stratification, and mechanistic insight require further investigation.
GRANTS
N. Yehya was supported by the Russell C. Raphaely Endowed Chair in Critical Care Medicine, Department of Anesthesiology and Critical Care Medicine, Children's Hospital of Philadelphia, and by
DISCLOSURES
Dr. Neal J. Thomas reports personal fees from Therabron and CareFusion, and grants from the FDA, all outside of the submitted work. Dr. Susan S. Margulies reports personal fees from Astrocyte Pharmaceuticals, outside of the submitted work. The remaining author declares no conflicts of interest.
AUTHOR CONTRIBUTIONS
N.Y., N.J.T., and S.S.M. conception and design of research; N.Y. performed experiments; N.Y. analyzed data; N.Y., N.J.T., and S.S.M. interpreted results of experiments; N.Y. prepared figures; N.Y. drafted manuscript; N.Y., N.J.T., and S.S.M. edited and revised manuscript; N.Y., N.J.T., and S.S.M. approved final version of manuscript.
REFERENCES
- 1.
The Acute Respiratory Distress Syndrome Network . Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med 342: 1301–1308, 2000.
Crossref | PubMed | ISI | Google Scholar - 2. . Circulating histones are mediators of trauma-associated lung injury. Am J Respir Crit Care Med 187: 160–169, 2013.
Crossref | PubMed | ISI | Google Scholar - 3. . The prognostic value of soluble intercellular adhesion molecule 1 plasma level in children with acute lung injury. J Intensive Care Med. In press.
ISI | Google Scholar - 4. . Circulating histones are major mediators of cardiac injury in patients with sepsis. Crit Care Med 43: 2094–2103, 2015.
Crossref | PubMed | ISI | Google Scholar - 5. . Circulating histone concentrations differentially affect the predominance of left or right ventricular dysfunction in critical illness. Crit Care Med 44: e278–288, 2015.
Crossref | ISI | Google Scholar - 6. . The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 149: 818–824, 1994.
Crossref | PubMed | ISI | Google Scholar - 7. . Circulating levels of KL-6 in acute respiratory distress syndrome sepsis or traumatic brain injury in critically ill children. Pediatr Pulmonol 41: 790–795, 2006.
Crossref | PubMed | ISI | Google Scholar - 8. . Association between C-reactive protein levels and outcome in acute lung injury in children. Eur J Pediatr 172: 1105–1110, 2013.
Crossref | PubMed | ISI | Google Scholar - 9. . Subphenotypes in acute respiratory distress syndrome: latent class analysis of data from two randomised controlled trials. Lancet Respir Med 2: 611–620, 2014.
Crossref | PubMed | ISI | Google Scholar - 10. . A Wilcoxon-type test for trend. Stat Med 4: 87–90, 1985.
Crossref | PubMed | ISI | Google Scholar - 11. . Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 44: 837–845, 1988.
Crossref | PubMed | ISI | Google Scholar - 12. . Impact of plasma histones in human sepsis and their contribution to cellular injury and inflammation. Crit Care 18: 543, 2014.
Crossref | PubMed | ISI | Google Scholar - 13. . Pediatric acute lung injury: prospective evaluation of risk factors associated with mortality. Am J Respir Crit Care Med 171: 995–1001, 2005.
Crossref | PubMed | ISI | Google Scholar - 14. . Early elevation of plasma soluble intercellular adhesion molecule-1 in pediatric acute lung injury identifies patients at increased risk of death and prolonged mechanical ventilation. Pediatr Crit Care Med 4: 315–321, 2003.
Crossref | PubMed | Google Scholar - 15. . Vasoactive-inotropic score as a predictor of morbidity and mortality in infants after cardiopulmonary bypass. Pediatr Crit Care Med 11: 234–238, 2010.
Crossref | PubMed | ISI | Google Scholar - 16. ,
International Consensus Conference on Pediatric Sepsis . International pediatric sepsis consensus conference: definitions for sepsis and organ dysfunction in pediatrics. Pediatr Crit Care Med 6: 2–8, 2005.
Crossref | PubMed | Google Scholar - 17. . Clinical relevance of circulating nucleosomes in cancer. Ann N Y Acad of Sci 1137: 180–189, 2008.
Crossref | PubMed | ISI | Google Scholar - 18. . Clinical use of circulating nucleosomes. Crit Rev Clin Lab Sci 46: 1–24, 2009.
Crossref | PubMed | ISI | Google Scholar - 19. . Unfractioned heparin for treatment of sepsis: A randomized clinical trial (The HETRASE Study). Crit Care Med 37: 1185–1196, 2009.
Crossref | PubMed | ISI | Google Scholar - 20. . Early elevation of matrix metalloproteinase-8 and −9 in pediatric ARDS is associated with an increased risk of prolonged mechanical ventilation. PLoS One 6: e22596, 2011.
Crossref | PubMed | ISI | Google Scholar - 21. . Nucleosomes in the pathogenesis of systemic lupus erythematosus. Rheum Dis Clin North Am 30: 529–558, 2004.
Crossref | PubMed | ISI | Google Scholar - 22. ,
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 40: 3238–3245, 2012.
Crossref | PubMed | ISI | Google Scholar - 23.
Pediatric Acute Lung Injury Consensus Conference Group . Pediatric acute respiratory distress syndrome: consensus recommendations from the Pediatric Acute Lung Injury Consensus Conference. Pediatr Crit Care Med 16: 428–439, 2015.
Crossref | PubMed | ISI | Google Scholar - 24. . Septic shock sera containing circulating histones induce dendritic cell-regulated necrosis in fatal septic shock patients. Crit Care Med 43: e107–116, 2015.
Crossref | PubMed | ISI | Google Scholar - 25. . Acute respiratory distress syndrome: the Berlin definition. JAMA 307: 2526–2533, 2012.
PubMed | ISI | Google Scholar - 26. . Drotrecogin alfa (activated) in adults with septic shock. N Engl J Med 366: 2055–2064, 2012.
Crossref | PubMed | ISI | Google Scholar - 27. . Early elevations in B-type natriuretic peptide levels are associated with poor clinical outcomes in pediatric acute lung injury. Pediatr Pulmonol 44: 1118–1124, 2009.
Crossref | PubMed | ISI | Google Scholar - 28. . Incidence and outcomes of acute lung injury. N Engl J Med 353: 1685–1693, 2005.
Crossref | PubMed | ISI | Google Scholar - 29. . Plasma soluble intercellular adhesion molecule-1 (sICAM-1) in pediatric ARDS during high frequency oscillatory ventilation: a predictor of mortality. Asian Pac J Allergy Immunol 23: 181–188, 2005.
PubMed | ISI | Google Scholar - 30. . Elevated PAI-1 is associated with poor clinical outcomes in pediatric patients with acute lung injury. Intensive Care Med 36: 157–163, 2010.
Crossref | PubMed | ISI | Google Scholar - 31. . Predictors of mortality in acute lung injury during the era of lung protective ventilation. Thorax 63: 994–998, 2008.
Crossref | PubMed | ISI | Google Scholar - 32. . Plasma biomarkers for acute respiratory distress syndrome: a systematic review and meta-analysis. Crit Care Med 42: 691–700, 2014.
Crossref | PubMed | ISI | Google Scholar - 33. . Oxygenation index predicts outcome in children with acute hypoxemic respiratory failure. Am J Respir Crit Care Med 172: 206–211, 2005.
Crossref | PubMed | ISI | Google Scholar - 34. . An early PEEP/FIO2 trial identifies different degrees of lung injury in patients with acute respiratory distress syndrome. Am J Respir Crit Care Med 176: 795–804, 2007.
Crossref | PubMed | ISI | Google Scholar - 35. . Prognostic determinants of acute respiratory distress syndrome in adults: impact on clinical trial design. Crit Care Med 33: S217–222, 2005.
Crossref | PubMed | Google Scholar - 36. . Prognostic and pathogenetic value of combining clinical and biochemical indices in patients with acute lung injury. Chest 137: 288–296, 2010.
Crossref | PubMed | ISI | Google Scholar - 37. . Postoperative course and hemodynamic profile after the arterial switch operation in neonates and infants A comparison of low-flow cardiopulmonary bypass and circulatory arrest. Circulation 92: 2226–2235, 1995.
Crossref | PubMed | ISI | Google Scholar - 38. . Extracellular histones are major mediators of death in sepsis. Nat Med 15: 1318–1321, 2009.
Crossref | PubMed | ISI | Google Scholar - 39. . Characterizing degree of lung injury in pediatric acute respiratory distress syndrome. Crit Care Med 43: 937–946, 2015.
Crossref | PubMed | ISI | Google Scholar - 40. . Circulating nucleosomes and severity of illness in children suffering from meningococcal sepsis treated with protein C. Crit Care Med 40: 3224–3229, 2012.
Crossref | PubMed | ISI | Google Scholar - 41. . Extracellular histones play an inflammatory role in acid aspiration-induced acute respiratory distress syndrome. Anesthesiology 122: 127–139, 2015.
Crossref | PubMed | ISI | Google Scholar - 42. . Incidence and outcomes of pediatric acute lung injury. Pediatrics 124: 87–95, 2009.
Crossref | PubMed | ISI | Google Scholar
