637937

research-article2016
PENXXX10.1177/0148607116637937Journal of Parenteral and Enteral NutritionWong et al

Original Communication
Journal of Parenteral and Enteral
Nutrition
Nutrition Delivery Affects Outcomes in Pediatric Volume XX Number X
Month 201X 1­–7
Acute Respiratory Distress Syndrome © 2016 American Society
for Parenteral and Enteral Nutrition
DOI: 10.1177/0148607116637937
jpen.sagepub.com
hosted at
online.sagepub.com
Judith Ju-Ming Wong, MBBCh BAO, MRCPCH1; Wee Meng Han, PhD2;
Rehena Sultana, MSc3; Tsee Foong Loh, MBBS, MRCPCH4;
and Jan Hau Lee, MBBS, MRCPCH, MCI4,5

Abstract
Background: Malnutrition is prevalent in critically ill children. We aim to describe nutrition received by children with acute respiratory
distress syndrome (ARDS) and to determine whether provision of adequate nutrition is associated with improved clinical outcomes.
Materials and Methods: We studied characteristics and outcomes of 2 groups of patients: (1) those who received adequate calories
(defined as ≥80% of predicted resting energy expenditure) and (2) those who received adequate protein (defined as ≥1.5g/kg/d of protein).
Outcomes of interest were mortality, ventilator-free days (VFDs), intensive care unit (ICU)–free days, multiorgan dysfunction, and need
for extracorporeal membrane oxygenation. Categorical variables were analyzed using the Fisher exact test, and continuous variables
were analyzed using the Mann-Whitney U test. Univariate and multivariate logistic regression models were used to identify associated
risk factors related to these outcomes of interest. Results: In total, 107 patients with ARDS were identified. There was a reduction in
ICU mortality in patients who received adequate calories (34.6% vs 60.5%, P = .025) and adequate protein (14.3% vs 60.2%, P = .002)
compared with those that did not. Patients with adequate protein intake also had more VFDs (median [interquartile range], 12 [3.0–19.0]
vs 0 [0.0–14.8] days; P = .005). After adjusting for severity of illness, adequate protein remained significantly associated with decreased
mortality (adjusted odds ratio [95% confidence interval], 0.09 [0.01–0.94]; P = .044). Conclusion: Our study demonstrated that adequate
nutrition delivery in children with ARDS was associated with improved clinical outcomes. Protein delivery may have potentially more
impact than overall caloric delivery. (JPEN J Parenter Enteral Nutr. XXXX;xx:xx-xx)

Keywords
acute respiratory distress syndrome; acute lung injury; nutrition; resting energy expenditure; pediatrics; pediatric intensive care unit;
critically ill children

Clinical Relevancy Statement
Studies demonstrate conflicting results regarding the impact of
adequate nutrition in adults with acute respiratory distress syn-
drome. The provision of enteral and parenteral nutrition in
children with acute respiratory distress syndrome, however,
has not been well documented. These patients are at significant
risk of nutrition deficiency and catabolism of lean body mass.
This study describes a high incidence of underfeeding and
details the amount and type of nutrition provided and the nutri-
tion adjuncts used. The association between adequacy of
caloric and protein requirements and improved clinical out-
comes (mortality and ventilator-free days) is demonstrated.
Introduction
Malnutrition is prevalent in up to 30%–45% of pediatric inten-
sive care unit (PICU) admissions.1–3 Among critically ill chil-
dren, malnutrition is associated with increased mortality,
multiorgan dysfunction, greater length of mechanical ventila-
tion (MV), and PICU stay.2–4 In patients with acute respiratory
distress syndrome (ARDS), the combination of established
malnutrition and severe respiratory insult theoretically leads to
a hypercatabolic state that renders the patient susceptible to
significant nutrition deficiencies, loss of lean body mass, and
deterioration of respiratory muscle strength.5
Adequate nutrient delivery is dependent on the accuracy of
nutrition assessment so as to avoid overfeeding or underfeeding
and the successful delivery of the prescribed nutrition.6–8 Certain
From the 1Department of Pediatrics, KK Women’s and Children’s
Hospital, Singapore; 2Department of Department of Nutrition &
Dietetics, KK Women’s and Children’s Hospital, Singapore; 3Centre
for Quantitative Medicine, Duke-NUS Medical School, The Academia,
169856, Singapore; 4Children’s Intensive Care Unit, Department of
Pediatric Subspecialities, KK Women’s and Children’s Hospital,
Singapore; and 5Office of Clinical Sciences, Duke-NUS School of
Medicine, The Academia, 169856, Singapore.
Financial disclosure: None declared.
Received for publication December 5, 2015; accepted for publication
February 12, 2016.
Corresponding Author:
Judith Ju-Ming Wong, MBBCh BAO, MRCPCH, Department of
Pediatrics, KK Women’s and Children’s Hospital, 100 Bukit Timah
Road, 229899, Singapore.
Email: Judith.wong.jm@kkh.com.sg

Downloaded from pen.sagepub.com at WESTERN MICHIGAN UNIVERSITY on March 28, 2016

2 Journal of Parenteral and Enteral Nutrition XX(X)
groups of patients (eg, sepsis, severe burns, trauma, and those
who have a prolonged PICU length of stay) more often develop
a hypermetabolic response and usually require more energy than
is estimated from predicted equations.9–11 Other patient groups
with traumatic brain injury and post–cardiac surgery may exhibit
hypometabolic responses.12,13 To our knowledge, no studies
have evaluated the energy expenditure in children with ARDS.
Data on nutrition practices, including the use of enteral
nutrition (EN), parenteral nutrition (PN), and adjuncts (eg,
motility agents, continuous feeds, transpyloric feeds), in pedi-
atric ARDS are also limited.14,15 To address this gap in the
medical literature, we conducted a retrospective cohort study
to describe the nutrition practices and delivery in pediatric
patients with ARDS within the first 7 days of illness. We
hypothesized that provision of adequate nutrition is associated
with improved clinical outcomes in children with ARDS.
Materials and Methods
Patients
KK Women’s and Children’s Hospital is 1 of 2 public tertiary
pediatric centers in Singapore. The PICU is a 16-bed multidis-
ciplinary facility. We generated a patient list from a search
using International Classification of Diseases and Snomed
Clinical Terminology codes as well as an independent search
of the PICU registry. All pediatric patients aged 0–18 years and
diagnosed with ARDS in their discharge or death summary
between January 1, 2009, and March 31, 2014, were included
in the analysis. The American-European Consensus Conference
(AECC) definition for ARDS was used.16 Premature neonates
(corrected age <35 weeks) and patients being cared for in the
neonatal intensive care unit were excluded. This study has
been approved by our hospital’s institutional review board, and
a waiver of consent was granted.
In our PICU, nutrition is prescribed by the managing physi-
cian with input from a PICU dietitian in certain cases (eg, mal-
nutrition, long stayers). The initiation of EN or PN is physician
dependent. In general, physicians prescribe standard polymeric
formula according to the Holliday Segar method.17 Dietitians
follow the American Society for Parenteral and Enteral
Nutrition (A.S.P.E.N.) guidelines for nutrition targets.18 Most
commonly, EN is commenced by intragastric bolus feeding
every 3–4 hourly. The use of nutrition adjuncts is dependent on
physician preferences.
Data Collection
We extracted demographic, clinical (eg, comorbidities, Pediatric
Index of Mortality 2 [PIM 2] score on admission, oxygenation
index [OI]), and chest x-ray information through chart review
using a predetermined data collection form. Data for OI were
collected after 24 hours of ARDS as this has been shown to be
more reflective of lung injury.19 We also extracted nutrition data
that included growth parameters, feeding volume, type of
Downloaded from pen.sagepub.com at WESTERN MICHIGAN UNIVERSITY on March 28, 2016
formula milk, use of additives, route of feeding, use of feeding
adjuncts, PN and EN orders, and macronutrient delivery. Our
PICU uses electronic nursing charts. Each hour, the actual
amount and type of feeds and PN received by the patient is
charted. Interruptions and delays are also captured. Our primary
outcome of interest is PICU mortality. Secondary outcomes
include ventilator-free days (VFDs), PICU-free days (IFDs),
multiorgan dysfunction, and need for extracorporeal membrane
oxygenation (ECMO).
Definitions
VFD is defined as days alive and free from MV up to 28 days
from the diagnosis of ARDS—this is to avoid concluding a short
duration of MV just because a patient died early in the disease
process.20 IFD was calculated as being alive and discharged from
the PICU up to 28 days from the diagnosis of ARDS. Multiorgan
dysfunction was defined to be present when 2 or more extrapul-
monary organ systems were involved using the International
Pediatric Sepsis Consensus Conference criteria.21 Resting energy
expenditure (REE) was estimated using the Schofield equation
without stress factor.4,22 Adequate calories were defined as
receiving at least 80% of REE by the third day of ARDS.
Adequate protein was defined as receiving at least 1.5 g/kg/d of
protein by the third day of ARDS.23 We defined overfeeding as
caloric delivery in excess of 120% of REE over any particular
day during the course of ARDS. We chose to use >120% because
the accuracy of predictive equations is known to be poorer than
indirect calorimetry (IC), and this portends greater variation in
predicted REE, especially for small children.24 Early EN was
defined as initiation of enteral feeding within 24 hours of ARDS.
Statistical Analysis
Demographic, clinical, nutrition, and survival outcomes are sum-
marized as frequency with corresponding proportion for categori-
cal variables and median (interquartile range [IQR]) for continuous
variables. Association between intermittent variables (adequate
calorie delivery, protein delivery, overfeeding, and early initiation
of EN) and other categorical variables was evaluated using the
Fisher exact test while association between continuous variables
(PIM 2 score, OI, median percentage, and median total calories
and protein) was evaluated using the Mann-Whitney U test.
We used univariate and multivariate logistic regression models
to quantify the association between potential confounding factors
and binary outcome variables of PICU mortality, multiorgan dys-
function, and ECMO. The association from logistic regression
was characterized using odds ratios (ORs) and corresponding
95% confidence intervals (95% CIs). All multivariate models
were adjusted for OI on day 2, PIM 2 score, presence of comor-
bidities, and whether adequate calories and protein were provided.
OI and PIM 2 score were considered continuous variables in the
multivariate analysis. Significance level was set at .05, and all
tests were 2-tailed. SAS version 9.3 software (SAS Institute, Cary,
NC) was used for the analysis.
Wong et al 3

Results
Of the 3057 PICU admissions between January 1, 2009, and
March 31, 2014, a total of 109 fulfilled the AECC criteria for
ARDS. Two patients were excluded from this study because
they were aged >18 years. Hence, a total of 107 patients (3.6%
of all PICU admissions) were analyzed. The median (IQR) age
was 5.2 (1.0–10.4) years, and weight was 16.0 (8.0–29.0) kg.
The overall median (IQR) PIM 2 score was 8.9% (3.7%–
21.7%) on admission, and the median (IQR) OI on the second
day of ARDS was 14.0 (7.8–25.4). Pneumonia (81/107; 75.7%)
was the main inciting event leading to ARDS in our cohort fol-
lowed by sepsis (45/107; 42.0%). Sixty-one (57.0%) patients
had an underlying chronic comorbidity. The 3 most common
comorbidities were neuromuscular (33.0%), hematology/
immunology (26.0%), and respiratory (16.0%). Overall PICU
mortality rate was 58 of 107 (54.2%). Figure 1.  Median percentage calories and median calories per
In total, 28 (26.2%) patients received early EN. The timeli- kilogram in survivors and nonsurvivors on days 1–7 of acute
ness of initiating EN was associated with decreased mortality respiratory distress syndrome. *Mann-Whitney test for median
(early vs late EN group: 26.2% vs 73.8%, P = .008). Twenty-six percentage calories, P < .05. +Mann-Whitney test for median
(24.3%) patients received PN. There was no difference between total calories/kg, P < .05. Median percentage calories—total
the median total calories received on the third day of ARDS calories delivered (kcal)/predicted resting energy expenditure
(kcal) × 100%. Median total calories/kg (kcal/kg).
between those with or without PN (18.6 [6.5–41.1] vs 9.3 [0.8–
37.6] kcal/kg, P = .311). The use of nutrition adjuncts was preva-
lent: 22 (20.6%) patients were taking motility agents
(domperidone or erythromycin), 81 (75.7%) were taking acid
suppressants (ranitidine or omeprazole), 42 (48.8%) were on
continuous feeding, and 5 (4.7%) were on transpyloric feeding.
Compared with patients on bolus feeds, the use of continuous
feeding was associated with a higher proportion of patients
achieving adequate protein (76.9% vs 43.8%, P = .036). Four
(3.7%) patients were taking nutrition additives (eg, beneprotein
and human milk fortifiers). Special formulas were used in 44
(41.1%) patients. These included high-calorie formulas (27;
25.2%) and other special formulas (eg, antireflux formulas, tran-
sitional formulas, hydrolyzed formulas) in 18 (16.8%) patients.
Median percentage calories delivered (calories as a percent-
age of estimated REE) was significantly higher in PICU survi-
vors compared with nonsurvivors for the first 5 days of ARDS
(Figure 1). Median total calories per kilogram (Figure 1) and
median total protein per kilogram (Figure 2) supplied during the Figure 2.  Median protein per kilogram delivered on days 1–7
course of ARDS were also significantly higher in PICU survi- of acute respiratory distress syndrome. *Mann-Whitney test for
median total protein/kg, P < .05. Median total protein/kg (g/kg).
vors for the first 7 days and 5 days of ARDS, respectively.
Twenty-six (24.3%) patients achieved adequate calories, and 14
(13.1%) achieved adequate protein intake by day 3 of ARDS Univariate logistic regression analysis demonstrated that the
(Tables 1 and 2). There was a significant reduction in PICU following risk factors influenced PICU mortality: inadequate
mortality in patients who received adequate calories compared caloric delivery, inadequate protein delivery, presence of comor-
with those with inadequate calories (36.4% vs 60.5%, P = .003) bidities, higher admission PIM 2 score, and higher OI on day 2 of
(Table 3). In patients who received adequate protein, PICU ARDS. Multivariate logistic regression analysis demonstrated
mortality was also lower compared with those who did not that only inadequate protein delivery, PIM 2, and oxygenation
receive adequate proteins (14.3% vs 60.2%, P = .002) (Table 4). index were independent predictors of increased ICU mortality
VFDs were significantly higher in patients who received ade- (Table S1). In the multivariate model, caloric and protein intake
quate protein compared with patients who did not (12.0 [3–19] were not demonstrated to be independent risk factors for decreased
vs 0 [0–4], P = .005). IFD, increased multiorgan dysfunction, and need for ECMO.

Downloaded from pen.sagepub.com at WESTERN MICHIGAN UNIVERSITY on March 28, 2016

4 Journal of Parenteral and Enteral Nutrition XX(X)

Table 1.  Characteristics of Patients Who Received Adequate and Inadequate Calories.a

Characteristic Inadequate Calories (n = 81) Adequate Calories (n = 26) P Value

Age, y .033
  <2 22 (27.2) 15 (57.7)
  2–5 12 (14.8) 4 (15.4)
  5–10 22 (27.2) 3 (11.5)
  >10 25 (30.9) 4 (15.4)
Weight, median (IQR), kg 18.0 (9.8–30.1) 9.4 (4.5–17.0) .006
Sex, male 38 (46.9) 9 (34.6) .364
Surgical admission 1 (1.2) 3 (11.5) .044
PIM 2 score, median (IQR) 9.8 (4–22) 6.3 (3.3–14.1) .201
OI on day 2, median (IQR) 13.8 (7.8–25.5) 15.0 (6.7–24.9) .693
Comorbidities 45 (55.6) 16 (61.5) .654
Prone position 20 (24.7) 12 (46.2) .049
Inotrope use 57 (70.4) 17 (65.4) .633
Paralysis 30 (37.0) 4 (15.4) .052
Need for hemodialysis 8 (9.9) 5 (19.2) .298
Length of stay, median (IQR), d 18.3 (7.1–43.6) 29.4 (12.2–51.7) .205

IQR, interquartile range; OI, oxygenation index = mean airway pressure × (fraction of inspired oxygen/partial pressure of arterial oxygen); PIM,
Pediatric Index of Mortality.
a
Adequate calories was defined as receiving at least 80% of resting energy expenditure by the third day of acute respiratory distress syndrome. Values are
presented as number (%) unless otherwise indicated.

Table 2.  Characteristics of Patients Who Received Adequate and Inadequate Protein.a

Characteristic Inadequate Protein (n = 93) Adequate Protein (n = 14) P Value

Age, y .088
  <2 28 (30.1) 9 (64.3)
  2–5 14 (15.1) 2 (14.3)
  5–10 24 (25.8) 1 (7.1)
  >10 27 (29.0) 2 (14.3)
Weight, median (IQR), kg 17.0 (9.6–30) 7.2 (4.0–14.0) .002
Sex, male 42 (45.2) 5 (35.7) .574
Comorbidities 53 (57.0) 8 (57.1) 1.000
PIM 2 score, median (IQR) 9.2 (4.0–21.7) 6.4 (3.3–14.1) .477
OI on day 2, median (IQR) 13.7 (7.4–25.4) 17.6 (11.2–24.8) .393
Prone position 22 (23.7) 10 (71.4) <.001
Inotrope use 66 (71.0) 8 (57.1) .355
Paralysis 30 (32.3) 4 (28.6) 1.000
Dialysis 12 (12.9) 1 (7.1) 1.000
Length of stay, median (IQR), d 17.2 (7.1, 42.6) 44.1 (29.4, 60.8) .005

IQR, interquartile range; OI, oxygenation index = mean airway pressure × (fraction of inspired oxygen/partial pressure of arterial oxygen); PIM,
Pediatric Index of Mortality.
a
Adequate protein was defined as receiving at least 1.5 g/kg/d or protein by the third day of acute respiratory distress syndrome. Values are presented as
number (%) unless otherwise indicated.

Within the first 7 days of ARDS, overfeeding occurred in 27 Discussion

(25.2%) patients with ARDS. PICU mortality was lower in the
group with overfeeding compared with the group that did not
have overfeeding (8 [29.6%] vs 50 [62.5%], P = .004). There
were more VFDs in the overfeeding group (7 [0–15] vs 0 [0–2]
days, P < .001). Extubation failure was not increased in patients
with overfeeding (16.0 vs 6.8%, P = .224) compared with
those who did not have overfeeding.
This study describes nutrient delivery in critically ill children
over the first 7 days of ARDS. Only a small proportion (26.1%)
of patients received any form of nutrition within the first 24
hours of ARDS. We observed nutrient delivery to be subopti-
mal. Only 24.2% and 13.0% of patients received adequate
calories and protein, respectively, by the third day. We

Downloaded from pen.sagepub.com at WESTERN MICHIGAN UNIVERSITY on March 28, 2016

Wong et al 5

Table 3.  Outcomes of Patients With Acute Respiratory Distress Syndrome Based on Adequacy of Caloric Intake.a

Clinical Outcome Inadequate Calories (n = 81) Adequate Calories (n = 26) P Value

PICU mortality 49 (60.5) 9 (34.6) .003
Ventilator-free days, median (IQR) 0 (0–4) 3 (0–12) .068
PICU-free days, median (IQR) 0 (0–15) 0 (0–17) .687
Multiorgan dysfunction 58 (72.5) 14 (53.8) .093
ECMO 6 (7.4) 1 (3.8) 1.000

ECMO, extracorporeal membrane oxygenation; IQR, interquartile range, PICU, pediatric intensive care unit.
a
Adequate calories was defined as receiving at least 80% of resting energy expenditure by the third day of acute respiratory distress syndrome. Values are
presented as number (%) unless otherwise indicated.

Table 4.  Outcomes of Patients With Acute Respiratory Distress Syndrome Based on Adequacy of Protein Intake.a

Clinical Outcome Inadequate Protein (n = 93) Adequate Protein (n = 14) P Value

PICU mortality 56 (60.2) 2 (14.3) .002
Ventilator-free days, median (IQR) 0 (0–4) 12.0 (3–19) .005
PICU-free days, median (IQR) 0.0 (0–15) 0.0 (0–14) .940
Multiorgan dysfunction 65 (70.7) 7 (50.0) .136
ECMO 5 (5.4) 2 (14.3) .227

ECMO, extracorporeal membrane oxygenation; IQR, interquartile range, PICU, pediatric intensive care unit.
a
Adequate protein was defined as receiving at least 1.5 g/kg/d of protein by the third day of acute respiratory distress syndrome. Values are presented as
number (%) unless otherwise indicated.

demonstrated an association between early and adequate nutri-
tion with lower PICU mortality and greater VFDs.
As critical illness is associated with increased basal meta-
bolic rate and protein catabolism, patients are at high risk of
developing malnutrition and loss of lean body mass that is
potentially detrimental to their recovery.1,2 Thus, adequate
energy and protein intake are important to buffer these stress-
induced changes.6,25 Our data are consistent with previous
observational studies demonstrating a reduction in PICU mor-
tality in patients with ARDS with improved caloric and protein
delivery. A multicenter study involving 500 intubated PICU
patients demonstrated that patients who received >66% of pre-
scribed calories via EN had lower 60-day mortality compared
with those who received <33% prescribed calories (OR, 0.14;
95% CI, 0.03–0.61; P = .002).3 Another observational study
showed a stepwise reduction in mortality with increasing pro-
tein delivery in critically ill adults.26 In children, a multicenter
study involving 1295 mechanically ventilated patients com-
pared mortality rates of children who received mean enteral
protein intake <20% and those with ≥60% of the prescribed
goal.27 The investigators found that protein adequacy was asso-
ciated with lower 60-day mortality (OR, 0.14; 95% CI, 0.04–
0.52; P = .003). Interestingly, our study demonstrated that after
adjusting for other prognostic factors, adequate protein but not
calories remains significantly associated with reduced mortal-
ity. We postulate that protein delivery may play a more impor-
tant role than caloric delivery in critical ill children.
In the general medical PICU population, overfeeding has
been reported to occur in up to 83% of patients.28 In our cohort
of patients with ARDS, overfeeding occurred in a quarter of
patients and was associated with improved PICU mortality and
VFDs. Previous studies using IC have used >110% of predicted
REE as the definition for overfeeding.4,28 We chose to use a
more conservative limit of >120% because the accuracy of pre-
dictive equations is known to be poorer than IC.24 A sensitivity
analysis using an overfeeding cutoff of >110% yielded similar
results. Patients in the overfeeding group had reduced mortality
(8 [29.6%] vs 50 [62.5%], P = .004) and increased VFDs (7
[0–15] vs 0 [0–2] days, P < .001) compared with those who
were not overfed. This positive effect contrasts with literature,
which reports negative associations, including azotemia, hyper-
tonic dehydration, metabolic acidosis (from protein overfeed-
ing), hyperglycemia, hypertriglyceridemia, hepatic steatosis
(from carbohydrate and fat overfeeding),29,30 and increased
work of breathing due to increased carbon dioxide production
leading to prolonged MV.31 These detrimental effects were
described in the adult population and have not been conclu-
sively reported in children. The association between overfeed-
ing and improved mortality and VFDs may indicate that we
were grossly underestimating the metabolic demands in chil-
dren. We studied this group further and found that survivors had
a smaller energy cumulative deficit compared with nonsurvi-
vors; however, this was not statistically significant (−1260.0
[−3722.0 to −280.0] vs −2138.0 [−4089.0 to −978.0] kcal, P =
.284). The maximum cumulative energy balance in our cohort
was +2788 kcal over the first 7 days of ARDS. These findings
suggest that the degree of overfeeding in this cohort was not
high. Moreover, we did not measure REE. Hence, the associa-
tion between overfeeding and improved outcomes should be
interpreted with caution, and further research is needed to more

Downloaded from pen.sagepub.com at WESTERN MICHIGAN UNIVERSITY on March 28, 2016

6 Journal of Parenteral and Enteral Nutrition XX(X)
accurately study the impact of overfeeding in critically ill
children.
The strength of this study is that we have identified all
patients with ARDS according to the AECC definition as all
PICU admissions were screened. The focus on a specific sub-
population of critically ill children reduces the bias of hetero-
geneity. ARDS literature on nutrition support in the pediatric
population is lacking, and our study contributes important data
on nutrition practices in this group of extremely ill children.
However, this study is not without limitations. The reasons for
not achieving adequate calories/protein and reasons for delay
in feed initiation and interruption were also not examined in
this study. Prescription of nutrition and the use of nutrition
adjuncts are dependent on physician preference. This nonstan-
dardization of practice may lead to bias, resulting in sicker
patients receiving less nutrition and delayed initiation of feeds.
We attempted to address this bias by adjusting for severity of
illness and other variables that were significant in the univari-
ate analysis. After multivariate adjustment, we still demon-
strated that protein adequacy was associated with mortality.
Another limitation is the small sample size. We performed a
power calculation to examine whether our sample size limited
the power of our study. We had 14 patients in the adequate pro-
tein group, with the proportion of nonsurvivors being 14.3% and
60.2% in the adequate and inadequate protein groups, respec-
tively. We found that the power of comparison was 85% based
on comparison of these 2 proportions using the Fisher exact test,
assuming α = 0.05 and a 1:3 allocation ratio. Similarly, we had
26 patients in the adequate calorie group, with the proportion of
nonsurvivors being 34.6% and 60.5% in the adequate and inad-
equate calorie groups, respectively, with a power of comparison
of 64.4%. Hence, the comparison between calorie groups was
underpowered, and this could be the reason that a significant
difference in survival was not demonstrated. IC is the gold-stan-
dard method of measuring REE, but its use in patients with
ARDS may be problematic and inaccurate. Technical factors
that decrease the accuracy of IC include fraction of inspired oxy-
gen that is high (>60%) or unstable, positive end-expiratory
pressure >12 cm water, hypo/hyperventilation, leak in the sam-
pling system, presence of chest tubes, the concurrent use of
dialysis, and hemodynamic and metabolic instability.32 All these
factors are common in patients with ARDS. Alternative methods
to measure energy expenditure such as doubly labeled water, cir-
culatory indirect calorimetry (Fick principle), and bioelectrical
impedance analysis have their own limitations.33 Given the limi-
tations of IC and these alternative methods, we adopted the use
of predictive equations instead in our study.
Conclusion
This study described nutrition delivery in critically ill children
with ARDS. We demonstrated that adequate calorie and pro-
tein delivery in children with ARDS was associated with
improved clinical outcomes. Attention should be given to pro-
vision of early and adequate nutrition in this group of patients.
Downloaded from pen.sagepub.com at WESTERN MICHIGAN UNIVERSITY on March 28, 2016
In addition, inadequate protein delivery was significantly asso-
ciated with PICU mortality, even after controlling for severity
of illness. Clinicians should consider optimizing protein deliv-
ery and not just total calories in children with ARDS. Due to
the limitations in the current study design, future prospective
trials should investigate the role of optimal nutrition in improv-
ing outcomes in pediatric ARDS.
Statement of Authorship
J. J.-M. Wong, W. M. Han, and J. H. Lee were involved in the con-
ception and design of the study, acquisition of data, interpretation of
data, and drafting, revising, and approval of the final manuscript. R.
Sultana was involved in the statistical analysis of data and critically
reviewed and approved the final manuscript. T. F. Loh was involved
in the interpretation of analysis and critically reviewed and approved
the final manuscript. J. J.-M. Wong and J. H. Lee are accountable
for all aspects of work ensuring integrity and accuracy. J. H. Lee
provided mentorship throughout the project.
Supplemental Material
Supplemental Table S1 is available online at http://pen.sagepub.com.
References
1. Prieto MB, Cid JL-H. Malnutrition in the critically ill child: the
importance of enteral nutrition. Int J Environ Res Public Health.
2011;8(11):4353-4366.
2. de Souza Menezes F, Leite HP, Koch Nogueira PC. Malnutrition as
an independent predictor of clinical outcome in critically ill children.
Nutrition. 2012;28(3):267-270.
3. Mehta NM, Bechard LJ, Cahill N, et al. Nutritional practices and their
relationship to clinical outcomes in critically ill children—an international
multicenter cohort study. Crit Care Med. 2012;40(7):2204-2211.
4. Briassoulis G, Venkataraman S, Thompson AE. Energy expenditure in
critically ill children. Crit Care Med. 2000;28(4):1166-1172.
5. Krzak A, Pleva M, Napolitano LM. Nutrition therapy for ALI and ARDS.
Crit Care Clin. 2011;27(3):647-659.
6. Skillman HE, Wischmeyer PE. Nutrition therapy in critically ill infants
and children. JPEN J Parenter Enteral Nutr. 2008;32(5):520-534.
7. Rogers EJ, Gilbertson HR, Heine RG, Henning R. Barriers to adequate
nutrition in critically ill children. Nutrition. 2003;19(10):865-868.
8. de Neef M, Geukers VG, Dral A, Lindeboom R, Sauerwein HP, Bos AP.
Nutritional goals, prescription and delivery in a pediatric intensive care
unit. Clin Nutr. 2008;27(1):65-71.
9. Suman OE, Mlcak RP, Chinkes DL, Herndon DN. Resting energy expen-
diture in severely burned children: analysis of agreement between indirect
calorimetry and prediction equations using the Bland-Altman method.
Burns. 2006;32(3):335-342.
10. Brandi LS, Santini L, Bertolini R, Malacarne P, Casagli S, Baraglia AM.
Energy expenditure and severity of injury and illness indices in multiple
trauma patients. Crit Care Med. 1999;27(12):2684-2689.
11. Tilden SJ, Watkins S, Tong TK, Jeevanandam M. Measured energy
expenditure in pediatric intensive care patients. Am J Dis Child.
1989;143(4):490-492.
12. Mtaweh H, Smith R, Kochanek PM, et al. Energy expenditure in chil-
dren after severe traumatic brain injury. Pediatr Crit Care Med.
2014;15(3):242-249.
13. Mehta NM, Costello JM, Bechard LJ, et al. Resting energy expenditure
after Fontan surgery in children with single-ventricle heart defects. JPEN
J Parenter Enteral Nutr. 2012;36(6):685-692.
14. Needham DM, Dinglas VD, Bienvenu OJ, et al. One year outcomes
in patients with acute lung injury randomised to initial trophic or full
Wong et al 7
enteral feeding: prospective follow-up of EDEN randomised trial. BMJ.
2013;346:f1532.
15. Rice TW, Wheeler AP, Thompson BT, et al. Initial trophic vs full enteral
feeding in patients with acute lung injury: the EDEN randomized trial.
JAMA. 2012;307(8):795-803.
16. 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(3, pt 1):818-824.
17. Meyers RS. Pediatric fluid and electrolyte therapy. J Pediatr Pharmacol
Ther. 2009;14(4):204-211.
18. Mehta NM, Compher C. A.S.P.E.N. Clinical guidelines: nutrition
support of the critically ill child. JPEN J Parenter Enteral Nutr.
2009;33(3):260-276.
19. Wong JJ, Loh TF, Testoni D, Yeo JG, Mok YH, Lee JH. Epidemiology of
pediatric acute respiratory distress syndrome in Singapore: risk factors and
predictive respiratory indices for mortality. Front Pediatr. 2014;2:78.
20. Schoenfeld DA, Bernard GR. Statistical evaluation of ventilator-free days
as an efficacy measure in clinical trials of treatments for acute respiratory
distress syndrome. Crit Care Med. 2002;30(8):1772-1777.
21. Goldstein B, Giroir B, Randolph A. International pediatric sepsis consen-
sus conference: definitions for sepsis and organ dysfunction in pediatrics.
Pediatr Crit Care Med. 2005;6(1):2-8.
22. Meyer R, Kulinskaya E, Briassoulis G, et al. The challenge of developing
a new predictive formula to estimate energy requirements in ventilated
critically ill children. Nutr Clin Pract. 2012;27(5):669-676.
23. Bechard LJ, Parrott JS, Mehta NM. Systematic review of the influence
of energy and protein intake on protein balance in critically ill children. J
Pediatr. 2012;161(2):333-339.e1.
Downloaded from pen.sagepub.com at WESTERN MICHIGAN UNIVERSITY on March 28, 2016
24. Framson CM, LeLeiko NS, Dallal GE, Roubenoff R, Snelling LK, Dwyer
JT. Energy expenditure in critically ill children. Pediatr Crit Care Med.
2007;8(3):264-267.
25. Lavery GG, Glover P. The metabolic and nutritional response to critical
illness. Curr Opin Crit Care. 2000;6(4):233-238.
26. Weijs PJ, Looijaard WG, Beishuizen A, Girbes AR, Oudemans-van
Straaten HM. Early high protein intake is associated with low mortality
and energy overfeeding with high mortality in non-septic mechanically
ventilated critically ill patients. Crit Care. 2014;18(6):701.
27. Mehta NM, Bechard LJ, Zurakowski D, Duggan CP, Heyland DK.
Adequate enteral protein intake is inversely associated with 60-d mortal-
ity in critically ill children: a multicenter, prospective, cohort study. Am J
Clin Nutr. 2015;102(1):199-206.
28. Mehta NM, Bechard LJ, Dolan M, Ariagno K, Jiang H, Duggan C. Energy
imbalance and the risk of overfeeding in critically ill children. Pediatr Crit
Care Med. 2011;12(4):398-405.
29. Klein CJ, Stanek GS, Wiles CE III. Overfeeding macronutrients
to critically ill adults: metabolic complications. J Am Diet Assoc.
1998;98(7):795-806.
30. Chwals WJ. Overfeeding the critically ill child: fact or fantasy? New
Horizons. 1994;2(2):147-155.
31. Askanazi J, Rosenbaum SH, Hyman AI, Silverberg PA, Milic-Emili J,
Kinney JM. Respiratory changes induced by the large glucose loads of
total parenteral nutrition. JAMA. 1980;243(14):1444-1447.
32. Schlein KM, Coulter SP. Best practices for determining resting energy
expenditure in critically ill adults. Nutr Clin Pract. 2014;29(1):44-55.
33. Pinheiro Volp AC, Esteves de Oliveira FC, Duarte Moreira Alves R,
Esteves EA, Bressan J. Energy expenditure: components and evaluation
methods. Nutr Hosp. 2011;26(3):430-440.