Journal of Parenteral and Enteral

Nutrition http://pen.sagepub.com/

Protocol-Driven Enteral Nutrition in Critically Ill Children: A Systematic Review

Judith Ju-Ming Wong, Chengsi Ong, Wee Meng Han and Jan Hau Lee
JPEN J Parenter Enteral Nutr 2014 38: 29 originally published online 26 September 2013
DOI: 10.1177/0148607113502811

The online version of this article can be found at:

http://pen.sagepub.com/content/38/1/29

Published by:

http://www.sagepublications.com

On behalf of:

The American Society for Parenteral & Enteral Nutrition

Additional services and information for Journal of Parenteral and Enteral Nutrition can be found at:

Email Alerts: http://pen.sagepub.com/cgi/alerts

Subscriptions: http://pen.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

>> Version of Record - Jan 2, 2014

OnlineFirst Version of Record - Sep 26, 2013

What is This?

Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014

502811
research-article2013
PENXXX10.1177/0148607113502811Journal of Parenteral and Enteral Nutrition XX(X)Wong et al

Review
Journal of Parenteral and Enteral
Nutrition
Protocol-Driven Enteral Nutrition in Critically Ill Children: Volume 38 Number 1
January 2014 2939
A Systematic Review 2013 American Society
for Parenteral and Enteral Nutrition
DOI: 10.1177/0148607113502811
jpen.sagepub.com
hosted at
online.sagepub.com
Judith Ju-Ming Wong, MBBCh BAO, MRCPCH1; Chengsi Ong, MS2;
Wee Meng Han, PhD2; and Jan Hau Lee, MBBS, MRCPCH3,4

Abstract
Enteral nutrition (EN) protocols are thought to improve clinical outcomes in the pediatric intensive care unit (PICU); however, critical
evaluation of their efficacy is limited. We conducted a systematic review with the aim of assessing the effect of EN protocols on
important clinical outcomes in these children. We searched MEDLINE, Cochrane Database for Reviews, Embase, and CINAHL using
predetermined keywords and MESH terms. We included randomized controlled trials (RCTs) and observational studies that involved
EN protocols in children admitted to the PICU for >24 hours. We included studies that reported at least 1 of our outcomes of interest.
Studies that exclusively studied premature neonates or adults were excluded. Primary outcomes were PICU or hospital mortality, PICU
or hospital length of stay (LOS), duration of mechanical ventilation, gastrointestinal (GI) complications, and infective complications.
Secondary outcomes were time to initiate feeds and time to achieve goal feeds. In total, we included 9 studies (total 1564 children) in our
systematic review (1 RCT, 4 before-and-after studies, 1 single-arm cohort study, 1 prospective descriptive study, and 2 audits). There is
low-level evidence that the use of EN protocols is associated with a reduction in GI and infective complications and improved timeliness
of feed initiation and achievement of goal feeds. Current medical literature does not have compelling data on the effects of an EN protocol
on clinical outcomes among critically ill children. Future clinical trials should look into using standardized interventions and outcome
measures to strengthen the existing evidence. (JPEN J Parenter Enteral Nutr. 2014;38:29-39)

Keywords
nutrition support; enteral nutrition; feeding algorithm; feeding protocol; critical care; children; pediatric intensive care unit

Introduction however, systematic evidence for improvement in clinical out-

Malnutrition is common among critically ill children and is
associated with increased morbidity and mortality.1-5 Compared
with parenteral nutrition (PN), enteral nutrition (EN) has been
demonstrated to be physiologically superior with fewer associ-
ated complications.6,7 In critically ill patients, EN has been
reported to (1) reduce risk of malnutrition, stress response,
severity and duration of catabolism, (2) preserve intestinal lining
integrity and gut flora, and (3) support host immune defenses.8-10
Despite this, the initiation of EN is often delayed due to respira-
tory instability, hemodynamic compromise, and the use of high
doses of sedation and neuromuscular blockade. Furthermore,
EN is frequently interrupted for intubation and extubation, bed-
side procedures, and radiological investigations.11-13
A Cochrane review in 2009 assessed the impact of early
feeding in critically ill children on clinical outcomes such as
mortality.14 The authors identified only 1 RCT that compared
early with late feeding, and concluded that there is insufficient
evidence to support early EN or PN in critically ill children and
therefore studies in this area are urgently needed.
Driven by reports of associated improvements in nutrient
delivery and ability to reach nutrition goals with protocol-
based feeding, a number of pediatric intensive care units
(PICUs) worldwide have started implementing EN protocols;15
comes such as mortality or length of stay (LOS) is lacking. We
conducted a systematic review to test our hypothesis that EN
protocols improve important clinical and nutrition outcomes in
critically ill children.
From 1Department of Pediatric Medicine, KK Womens and Childrens
Hospital, Singapore; 2Department of Nutrition and Dietetics, KK
Womens and Childrens Hospital, Singapore; 3Childrens Intensive
Care Unit, Department of Pediatric Subspecialties, KK Womens and
Childrens Hospital, Singapore; and 4Office of Clinical Sciences, Duke-
NUS Graduate School of Medicine, Singapore.
Supplementary material for this article is available on the Journal of
Parenteral and Enteral Nutrition website at http://jpen.sagepub.com/
supplemental.
Financial disclosure: None declared.
Received for publication May 29, 2013; accepted for publication August
2, 2013.
This article originally appeared online on September 26, 2013.
Corresponding Author:
Judith Ju-Ming Wong, MBBCh BAO, MRCPCH, Department of
Pediatric Medicine, KK Womens and Childrens Hospital, 100, Bukit
Timah Road, Singapore 229899, Singapore.
Email: judith.wjm@gmail.com.

Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014

30 Journal of Parenteral and Enteral Nutrition 38(1)
Methods
Selection of Studies
We defined the review question using the PICOTS (patient,
intervention, comparator, outcome, time, and setting) frame-
work. We included RCTs and observational studies to get a bet-
ter overview of the medical literature on overall effectiveness
of feeding protocols in critically ill children. Two authors inde-
pendently carried out an online search in MEDLINE, Cochrane
Database for Reviews, Embase and CINAHL using predeter-
mined keywords and MESH terms. The period of review was
from 1979 to October 2012. Keywords used included EN,
feeding, nutrition support, protocol, algorithm, guideline, pedi-
atric intensive care unit (PICU), and critical illness (details of
our search strategy for MEDLINE is outlined in Electronic
Appendix A).
Selection of Participants
Participants were patients between 1 day and 16 years of age
who have been admitted to a PICU for >24 hours. We accepted
admissions to all types of PICUs (medical, surgical, cardiotho-
racic, neurosurgical and oncology). Studies were excluded if
participants were primarily premature infants (gestational age
<35 weeks) admitted in a neonatal ICU. We excluded prema-
ture infants because these patients constitute a very different
population from critically ill children with special consider-
ations that could impact the progression of feeds (eg, reduced
gut maturation and risk of necrotizing enterocolitis [NEC]).16
Types of Interventions and Outcome
Measures
The intervention group was defined as patients on an EN pro-
tocol. The control group, if any, was defined as patients on the
existing feeding practices or the standard of care before the
introduction of any feeding protocol in their respective units
(ie, nonprotocolized feeding). We anticipated that EN proto-
cols may differ from one center to another (eg, utilization of
different routes of administration [nasogastric, nasoduodenal,
nasojejunal] and use of motility agents), but we accepted this
variation and considered these protocols collectively as the
intervention of interest.
Studies that reported any of the following primary outcomes
were considered: PICU/hospital mortality, PICU/hospital LOS,
duration of mechanical ventilation, gastrointestinal (GI) com-
plications/intolerances, and infective complications. GI compli-
cations included vomiting, diarrhea, abdominal distension and
any other GI concerns that required cessation of feeds. The need
to add on agents that reduced gastric acid secretion (eg, hista-
mine-2 antagonists, proton pump inhibitors, antacids) was not
considered a GI complication. In addition, we included studies
reporting our secondary outcomes of interest: time to initiate
feeds and time to achieve goal feeds or calories.
Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014
Data Extraction and Management
Two authors independently carried out a search in MEDLINE,
Cochrane Database for Reviews, Embase, and CINAHL using
predetermined keywords and MESH terms. All retrieved arti-
cles were screened for relevance based on title and abstract.
We then obtained the full text of shortlisted articles. Authors
independently decided on inclusion of studies and any dis-
crepancies were resolved through discussion after detailed
reexamination of the articles. We extracted data from each
study using standardized data extraction forms and all forms
were counterchecked for accuracy.
We applied the Grading of Recommendations Assessment,
Development and Evaluation (GRADE) assessment method
to the studies included to determine the quality of evidence
across outcomes. The GRADE method considers risk of bias,
directness of evidence, heterogeneity, precision of effect and
risk of publication bias as factors determining the quality of
evidence. This is a validated method to uniformly rate evi-
dence and recommendations and is widely endorsed by many
organizations (eg, Agency for Healthcare Research and
Quality, United States, National Institute for Clinical
Excellence, United Kingdom, Cochrane Collaboration and
World Health Organization).17-19
Results
We identified 876 abstracts from our search strategy (Figure 1).
Of these, 9 studies were relevant to our analysis1 RCT and 8
observational studies (see Supplemental Table 1).
Gottschlich et al studied children with acute burns and was
the only RCT that utilized an EN protocol.20 The objective of
this study was to evaluate the benefit of early (within 24 hours)
vs late (after 48 hours) initiation of enteral feeds while using
the same feeding protocol in both instances. Although this
study did not find significant effects of early feeding on out-
comes such as mortality and LOS, their results were not repre-
sentative of protocolized vs nonprotocolized feeding. Two
studies (n = 161) involved infants with hypoplastic left heart
syndrome (HLHS) who have undergone stage 1 palliation with
a Norwood procedure,21,22 the other studies involved a multi-
disciplinary cohort of patients (n = 1235).15,23-27
Among the observational studies, 4 studies were before-
and-after studies that involved a retrospective or historic
comparison group.21-24 One study, conducted by Mehta et al,
was an observational multicenter study of feeding practices
and contained a subgroup analysis comparing PICUs using a
feeding protocol with PICUs not using a feeding protocol.15
One study, by Briassoulis et al, was a single arm prospective
study that looked at the safety and efficacy of early and
aggressive feeding using a feeding protocol.27 The 2 remain-
ing studies, by Meyer et al and Tume et al, were audits con-
ducted to study the compliance and effect of a feeding
protocol in the PICU. Meyer et al evaluated baseline nutrition
Wong et al 31

PubMed Cochrane Database Embase CINAHL

1979-15th Oct 2012 of Systemac 1990-15th Oct 2012 1993-15th Oct 2012
159 citaons Reviews 427 citaons 189 citaons
1982-15th Oct 2012
101 citaons

854 non duplicate

citaons screened
742 citaons excluded
aer screening the tle
and abstract

112 citaons
retrieved

46 were quesonnaire/
surveys, reviews, clinical
pracce guidelines, leers
or commentaries, or
teaching modules

36 studies involved
exclusively adults or
premature neonate

21 did not ulise feeding

protocols or did not report
any of the outcomesof
interest

9 studies included

Figure 1. PRISMA (preferred reporting items for systematic reviews and meta-analyses) flowchart showing results of the systematic
search for randomized controlled trials and observational studies that involved feeding protocols for enteral nutrition in children (1 day to
16 years of age) admitted to the pediatric intensive care unit.

practices and compared 4 different audit periods during
which the feeding protocols used were introduced and subse-
quently improved.25 Tume et al conducted an audit and did
not have a control group for comparison.26
Effects of Interventions
Outcomes of interest contributed by each study are summa-
rized in Table 1. After applying the GRADE method, the level
of evidence for each outcome is summarized in Table 2.17
PICU length of stay and hospital length of stay, PICU mortal-
ity and hospital mortality. None of the 4 studies that examined
PICU LOS20,22-24 or the 3 that examined hospital LOS20-22,24
found significant differences between the EN protocol and the
control group. Four studies20,21,23,27 found no difference in
PICU mortality. Hospital mortality was not measured by any of
the included studies.
Duration of mechanical ventilation, infective complications,
and GI complications. Three studies found no difference in

Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014

32 Journal of Parenteral and Enteral Nutrition 38(1)

Table 1. Summary of the Outcomes in the 9 Studies Included in This Systematic Review.

Outcome of Interest Study Effect of EN Protocol

24
PICU length of stay Petrillo-Albarano et al
Braudis et al22
Geukers et al23
Gottschlich et al20
Hospital length of stay Petrillo-Albarano et al24
Braudis et al22
del Castillo et al21
PICU mortality Briassoulis et al27
Geukers et al23
del Castillo et al21
Gottschlich et al20
Hospital mortality Not measured
Duration of mechanical Braudis et al22
ventilation Geukers et al23
Gottschlich et al20
GI complications Petrillo-Albarano et al24
Braudis et al22
del Castillo et al21
Geukers et al23
Gottschlich et al20
Infective complications del Castillo et al21
Mehta et al15
Gottschlich et al20
Time to initiate feeds Braudis et al22
Meyer et al25
del Castillo et al21
Geukers et al23
Gottschlich et al20
Time to achieve goal Petrillo-Albarano et al24
feeds Braudis et al22
del Castillo et al21
Geukers et al23
Gottschlich et al20
No difference (mean 12.5 vs 12.9 days; P = .60)
No difference (mean 10 vs 12 days; P = .07)
No difference (median 8 vs 8 days; P = .54)
No difference (mean 54.8 5.9 vs 54.8 4.6 days; P = .95)a
No difference (mean 20.8 vs 22.6 days; P = .40)
No difference (mean 16 vs 19 days; P = .07)
No difference (median 21.5 vs 28 days; P = .25)
Overall mortality was 4 patients (5.6%)b
No difference (7.1% vs 8%; P = .82)
No difference (11% vs 19%; P = .27)
No difference (11% vs 8%; P = .99)a
No difference (mean 6 vs 7 days;, P = .19)
No difference (median 6.5 vs 6.8 days; P = .96)
No difference (mean 24.5 4.6 vs 22.5 4.2 days; P = .74)a
Less likely to experience diarrhea (2% vs 12%; P = .009) with relative riskc of
0.18 [CI: 0.04, 0.78], P = .02
Less constipation (33% vs 51%; P = .012)
No significant difference in incidence of NEC (0% vs 11%; P = .07)
Incidence of NEC was lower (6.5% vs 27%; P < .01) with relative riskc of 0.24
[0.07-0.79], P = .02. The effect was similar when comparing NEC in patients
who underwent mBTS in both groups (0% vs 27%; P = .007)
Reduced vomiting (0.7% vs 2.1%; P = .05)
No difference in incidence of diarrhea (0.8% vs 1.4%; P = .29) or gastric
residuals (4.1% vs 6.3%; P = .10)
No difference in diarrhea (63% vs 58%; P = .62)a
No difference in bowel necrosis (11% vs 3%; P = .35)a
No difference in incidence of positive blood cultures (15% vs 21%; P = .6)
Patients admitted to units which utilized a feeding protocol had a lower incidence
of acquired infections (odds ratio 0.18 [0.05, 0.64]; P = .008), and this was
independent of the severity of illness and the amount of energy or protein
intake
No difference in incidence of clinical sepsis (47% vs 58%; P = .23)a
No significant difference (median 4 vs 4 days; P = .85)
Median time for initiating nutrition support was reduced with protocol; 15 hrs in
1994-1995; 8 hrs in 1997-1998; 5.5 hrs in 2001 and 4.5 hrs in 2005. Median
reduction in time between 1994 and 2005 was statistically significant; P < .001
Longer to start feeds (median 7.5 vs 5.5 days; P < .001)
Higher percentage of patients on feeds on day 1 (78% vs 40%; P < .01) and on
day 2 (92% vs 60%, P < .01)
Started feeds earlier (mean 15.6 1 vs 48.5 0.4 hrs; P < .0001)a
Goal feeds reached sooner (mean 18.5 vs 57.8 hrs; P < .0001)
Goal feeds reached sooner (median 9 vs 13 days; P = .01)
Longer to reach goal feeds (median 7 vs 4 days; P = .02)d
Achieved 85% of the mean energy goal earlier (day 3 vs day 4; p <0.05)
The control group showed significant cumulative caloric deficit compared with
the intervention group in the first week post burn (mean -1960 181 kcal
balance; P < .0001) as well as in the and second week (mean -2734 360 vs
-898 449; P = .0022)a

CI, confidence interval; EN, enteral nutrition; GI, gastrointestinal; hrs, hours; mBTS, modified Blalock-Taussig shunt; NEC, necrotizing enterocolitis;
PICU, pediatric intensive care unit; RDA, recommended dietary allowance.
a
The aim of this study was to compare early vs late feeding using the same EN protocol in both arms.
b
This study did not involve a control group.
c
Significant relative risk indicating strong evidence of association.
d
The study authors suggested that this might have been due to more careful monitoring of feed intolerance on protocol.

Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014

Wong et al 33

Table 2. Summary of Findings Showing Level of Evidence Across Outcomes.

Outcome Result Level of Evidencea

PICU length of stay Results not pooled Low
No difference
Hospital length of stay Results not pooled Low
No difference
PICU mortality Results not pooled Very lowb
No difference
Hospital mortality Not measured
Duration of ventilation Results not pooled Low
No difference
Infective complications Results not pooled Low
Reduced infection
Gastrointestinal complications Results not pooled Lowc
Reduced GI complications
Time to start feeds Results not pooled Low
Improves time to start feeds
Time to achieve goal feeds Results not pooled Low
Improves time to achieve goal feeds

GI, gastrointestinal; PICU, pediatric intensive care unit.

a
Based on Grading of Recommendations Assessment, Development and Evaluation (GRADE).
b
Briassoulis et al27 compared mortality figures to admission severity scores and was inconsistent with 2 other studies that showed no difference.
c
We did not upgrade the level of evidence for this outcome despite significant relative risk reduction given the threat to validity in report by del Castillo
et al21 in which a change in surgical practice was instituted midway through the study and inconsistent findings in study by Braudis et al.22

duration of mechanical ventilation between groups.20,22,23
Three studies15,20,21 explored infective complications (eg,
blood stream infection, ventilator associated pneumonia, uri-
nary tract infections and wound sepsis.) After accounting for
severity of illness and nutrition intake variation, Mehta et al
reported that units utilizing a feeding protocol had lower rates
of acquired infections (odds ratio [OR] 0.18, 95% confidence
interval [CI]: 0.05, 0.64). Del Castillo et al and Gottschlich et
al, however, found no difference in infective complications
between groups.
Five studies explored the rate of GI complications.20-24
Petrillo-Albarano et al found a significant reduction in diarrhea
and constipation after the feeding protocol was implemented,
but it is worthwhile to note that their protocol included an
aggressive bowel regimen. In contrast, Geukers et al found sig-
nificantly lower rates of vomiting but not diarrhea. In infants
with HLHS, del Castillo et al found a significant reduction in
incidence of NEC after the implementation of a feeding proto-
col, while Braudis et al observed a trend toward reduction (0%
vs 11%; P = .07) but did not reach a statistically significant
difference. An important consideration in the interpretation of
these results is that midway through both studies there was a
change in surgical practice. In the study by Braudis et al, 33%
of patients in the study group underwent a Norwood procedure
with modified Blalock-Taussig Shunt (mBTS) while the
remaining patients underwent the procedure with right
ventricle to pulmonary artery conduit (Sano modification).
This is in contrast with the control group where all patients had
mBTS insertion. There was also a difference in the proportion
of patients undergoing each surgical technique between the
control and EN protocol group in the study by del Castillo et al.
The change in surgical techniques could have affected the out-
comes of these studies.28,29
Time to initiate feeds and time to achieve goal feeds. Two
studies found improvement in time to start feeding,23,25 1
showed no difference22 and 1 found it took longer to start
feeds while on an EN protocol.21 Three of 4 studies found
that groups on a feeding protocol reached their goal feeds
faster.22-24 The study by del Castillo et al however, found that
the protocol group reached goal feeds 3 days later than the
pre-protocol group (7 vs 4 days, P < .05). Causes of interrup-
tion cited by the studies included (1) intubation or extuba-
tion, (2) medical, surgical or radiological procedures, (3)
hemodynamic instability, (4) feed intolerance and, (5) fluid
restriction.
GRADE recommendation. Using the GRADE assessment
method, there is low level evidence that the use of an EN pro-
tocol reduces GI complications and infective complications,
and improves time to initiate feeds and time to achieve goal
feeds (Table 2).

Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014

34 Journal of Parenteral and Enteral Nutrition 38(1)
Discussion
Our systematic review on the use of EN protocols in PICUs
showed that the introduction of an EN protocol may (1)
enhance feed tolerance through reduction of GI complica-
tions, (2) reduce infective complications, and (3) improve
timeliness of feed initiation and achievement of goal feeds in
critically ill children. However, there is low-level evidence
that EN protocols make no difference to PICU/hospital mor-
tality, PICU/hospital LOS, and duration of mechanical venti-
lation. This is likely due in part to the small sample size and
lack of RCTs in the pediatric population and implies that fur-
ther research is very likely to change the estimate of effect for
these outcomes.
In this section, we discuss the (1) advantages and properties
of a successful EN protocol, (2) evidence from the adult popu-
lation, (3) limitations of pediatric studies (the 3 main areas of
limitations include study methodologies, heterogeneity in
patient populations and protocols, and outcome measures), and
finally (4) propose future directions for pediatric nutrition
studies.
Evidence-based protocols have several advantages: they
promote effective interventions, discourage ineffective ones,
and improve cost effectiveness, interdisciplinary cooperation
and consistency.30 Protocols also serve as an effective training
tool for the junior doctors, nurses and other support staff.31
Over the past years, introduction of protocols in various aspects
of critical care management in the PICU (eg, weaning from
mechanical ventilation, glycemic control, sedation and analge-
sia) have led to better outcomes.32-34
Protocols should incorporate best available evidence into a
clear, concise and user friendly framework. There should be a
transparent description of the process used for its development
and it should be evidence-based. For a protocol to be success-
ful, cultural differences and available resources need to be con-
sidered. EN protocols aim to automate feed initiation and
advancement in a safe manner, reduce the number and duration
of cessation of feeds, and remove barriers to optimal EN deliv-
ery.35 EN protocols should also delineate strategies to trouble-
shoot feed intolerance. These advantages make implementation
of EN protocols desirable.
Our systematic review highlights the scarcity of pediatric
literature. In contrast, evidence of protocol feeding is more
robust in critically ill adults. Multicenter RCTs have demon-
strated that EN protocols are nutritionally beneficial and asso-
ciated with improved clinical outcomes in this group of
patients.36-38 Martin et al36 showed that patients from ICUs (n =
499) that used feeding protocols received significantly more
EN (6.7 vs 5.4 days of EN per 10 patient-days; P = .042), and
a significantly shorter hospital stay (mean 25 vs 35 days; P =
.003). After adjusting for severity of illness and baseline nutri-
tion status, Barr et al37 (n = 200) showed that patients in the EN
Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014
protocol group were fed more frequently via the enteral route
(OR 2.4; 95% CI: 1.2-5.0; P = .009). The duration of mechani-
cal ventilation was also shown to be shorter (mean 17.9 31.3
vs 11.2 19.5 days; P = .03). Of note, the risk of death was
56% lower in patients who received EN (hazard ratio 0.44;
95% CI: 0.24, 0.80; P = .007). In comparison, our systematic
review demonstrated no significant reduction in PICU mortal-
ity or LOS among critically ill children. The lack of change in
PICU mortality is likely due to the lower overall mortality rates
in PICUs compared with adult ICUs.15,38 Notably, none of our
included studies reported negative clinical outcomes, suggest-
ing a low risk of adverse effects when using EN protocols.
EN protocols can also help improve timeliness of feed ini-
tiation. Doig et al38 (n = 1118) showed that ICUs using an EN
protocol started EN for patients earlier (mean 0.75 vs 1.37 days
with mean difference of 0.62 [95% CI: 0.82, 0.36; p <
.001]), started PN slightly earlier (mean 1.04 vs 1.40 days with
mean difference of 0.35 [95% CI: 0.61, 0.01; P = .04]) and
achieved caloric goals more often (6.10 vs 5.02 days per 10 fed
patient-days with mean difference of 1.07 [95% CI: 0.12, 2.22;
P = .03]). In our review, we found mixed results, with 3 of 4
studies showing a reduction in time to achieve goal feeds.22-24
The study by del Castillo et al, however, showed that it took
longer to start and achieve goal feeds while on an EN protocol.
The authors suggested that this might have been due to more
careful monitoring of feed intolerance on protocol. Even with
delayed feeding, the protocol group had a trend toward shorter
hospital stay, which they considered an improved clinical
outcome.
The use of EN protocols have also been shown to help over-
come feed delays due to GI intolerance. Woien et al39 found
that EN protocols in adults resulted in greater consistency in
nursing practices with respect to aspiration of gastric contents
and rate of increment in enteral feeding. Interruptions in EN
support was closely related to the lack of aspiration routines,
and this led to prescribed nutrition goals not being delivered to
the patient. Among the studies included in this review, there
was no difference in gastric residual volume (GRV) between
protocol and nonprotocol group, yet the protocol group had
less vomiting and received more feeds.23 Heyland et al35 ana-
lyzed 2 international observational cohort studies conducted in
2007 and 2008 (involving 269 centers with 5497 patients)
looking at the impact of EN protocols on nutrition delivery in
adult ICUs. 77% of the sites utilized a standardized EN proto-
col. The study found that sites utilizing an EN protocol had
better nutrition practices. The improvements included earlier
EN, greater overall caloric adequacy and greater calories from
EN, and greater use of motility agents in patients with high
GRVs. Evidence from adult studies demonstrates that EN pro-
tocols can optimize nutrition delivery, minimize the risks of
EN and improve clinical outcomes. Although pediatric obser-
vational studies suggested similar outcomes, rigorous research
Wong et al 35
and recommendations in children are still lacking.14 Our sys-
tematic review revealed limitations in the available studies
examining the effectiveness of EN protocols in critically ill
children. These limitations should be examined so that quality
of future studies can be improved and optimized.
Current available pediatric studies involve a relatively
small number of patients and have limitations in their method-
ologies. Four before-and-after studies21-24 may be prone to
performance bias. The EN protocol group being studied at a
later date may have benefited from other new medical inter-
ventions or changes in physician experience. The lack of a
comparison group in Briassoulis et als single arm prospective
cohort makes it difficult to come to a definitive conclusion that
EN protocols led to improved outcomes.27 The only RCT that
was included in our review did not specifically address the
comparison between protocol feeding vs nonprotocol feed-
ing.20 The limited number of RCTs highlights the unique chal-
lenges in the conduct of clinical trials in pediatric critical care
nutrition.40 Children may have preexisting under- or overnutri-
tion states, may have different growth trajectories depending
on their age, and may be chronically or acutely ill. Until more
rigorous studies are conducted, clinicians may need to consider
allocating greater consideration to observational studies in
nutrition care of the critically ill child. Perhaps one of the more
encouraging indications for promoting protocol feeding comes
from the descriptive study by Mehta et al.15 This multicenter
study found that the use of an EN protocol was associated with
decreased infective complications and improved survival in
children who received more than one-third of prescribed feed-
ing amount.
A majority of the studies did not have detailed information
on severity of illness, case mix of patients, or timing of nutri-
tion intervention. Patient population heterogeneity and selec-
tion (Table 3), and complexity of nutrition interventions may
influence outcomes across these studies. Protocol heterogene-
ity was also present among studies (Table 4). This is not sur-
prising because previous evidence had demonstrated that EN
protocols vary widely across centers.35 We attempted to exam-
ine the EN protocols in detail but 3 protocols were not avail-
able even after attempts to contact the authors.15,20,25 Although
EN protocols vary, its use has led to improvement in efficiency
and adequacy of nutrient delivery, less use of the PN route and
reduced interruptions to EN.35,39,41 Last, significant changes in
clinical outcomes (eg, mortality and LOS) are difficult to dem-
onstrate in children. Nevertheless, these outcomes were inves-
tigated and reported in the medical literature. We considered
Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014
including nutrition outcomes (eg, anthropometry, nutrition bio-
markers like serum albumin level and preserum albumin
level, and muscle function tests) in the planning stages of our
review. However, we focused on clinical outcomes because
improvement in surrogate outcomes do not always lead to
improved clinical outcomes.42
Future studies should be conducted to further clarify the
effectiveness of the various elements within an EN protocol.
These elements include the role of early vs late use of feed-
ing protocols, gastric vs postpyloric feeding as first line
route, and the utility of GRVs in children.20,44,45 Without
these studies, protocols will continue to be heterogeneous
and future reviews and individual studies may fall short of
their goal to demonstrate benefit. Only when the appropriate
protocol elements have been identified based on evidence,
larger and preferably multicenter studies like those con-
ducted in the adult population should be undertaken. Future
research studies should also target specific populations (eg,
post cardiac surgery or sepsis) and specific age groups (eg,
<2 years vs >2 years)8,46 to better define the benefits and
identify specific nutrition strategies for each group. Studies
and reviews in this field should also consider long-term out-
comes such as growth or intellectual delay, developmental or
behavioral outcomes, and development of critical illness-
related metabolic disease.
Taking into account our limitations and based on the
GRADE method, we can make only a weak recommendation
for the use of an EN protocol in critically ill children. However,
feeding protocols are associated with low risks and none of the
included studies reported negative outcomes. Moreover, the
use of an EN protocol can reduce the use of PN, which in turn
can reduce cost and risk of infective complications.47 In line
with recommendations from other investigators, we strongly
concur that the primary objective of a feeding protocol is to
ensure achievement of nutrient goals in a safe and timely
fashion.35,48
Conclusion
Delivery of nutrition in critically ill children should be opti-
mized. However, current medical literature shows only a low
level of evidence that the use of an EN protocol reduces GI and
infective complications, and improves time to initiate feeds
and time to achieve goal feeds. There is pressing need for fur-
ther studies to establish more favorable nutrition practices and
define their clinical benefits.
 36
Table 3. Characteristics of Patients and PICUs in Included Studies.
del Castillo et al21 (n = 98) Geukers et al23 (n = 171)
(Median, IQR) (Median, IQR)
Petrillo-
Mehta et al15 Albarano et al24 Braudis et al22 Meyer et al25 Tume et al26 Briassoulis Gottschlich
Characteristicsa (n = 500) (n = 184) (n = 63) (n = 353) (n = 47) et al27 (n = 71) et al20 (n = 77) Pre-protocol Post-protocol Pre-protocol Post-protocol
b
Age 4.5 (5.1) yrs 61.9 (72.9) mths 2 (1.7) weeks 3.3 (4.2) yrs 10 (42.0) mths 71.8 (57.2) mths 9.4 (4.2) yrs 6 (3, 9.5) days 6 (3, 10) days 4.7 (0.9, 18.9) 6.6 (0.9, 30.1)
mths mths
Weight (kg)b 20.3 (21.7) NS 3.3 (0.6) NS 8.4 (13.3) 24.9 (16.0) 43.3 (23.9) NS 6.1 (3.6, 11.8) 7.2 (3.8, 14.0)
Severity indexb PRISM III NS PRISM III NS NS PRISM NS PRISM II PRISM II PIM 2 3.5 PIM 2 4.2
9.6 (8.0) 15.3 (4.8) 12.8 (7.1) 32 (25.5, 39.5) 32 (25, 36) (4.4, 2.0) (5.1, 3.1)
PIM 6.4 (14.5)
Type of PICU(s) M M Cardiac M M M Burns Cardiac M
Include surgical Yes No Yes Yes Yes Yes Yes Yes Yes
patients

IQR, interquartile range; M, multidisciplinary; mths, months; NS, not specified; PICU, pediatric intensive care unit; PIM, pediatric index of mortality; PRISM, pediatric risk of mortality; yrs, years.
a
Continuous variables are expressed as means and standard deviations unless specified otherwise.
b
Where possible, values are expressed as mean with standard deviations or medians with interquartile ranges. Where appropriate, standard errors are converted to standard deviations. In studies that
provided data as median and ranges, means and standard deviations are estimated.43

Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014

 Table 4. Characteristics of EN Protocols in Included Studies (n = 6).
Characteristics of
protocols Briassoulis et al27 Petrillo-Albarano et al24 Braudis et al22 Tume et al del Castillo et al21 Geukers et al23

Inclusion/exclusion NS Clear exclusion criteria NS NS Clear exclusion criteria Clear exclusion criteria
criteria given given given
Indication for PN NS NS If EN not tolerated If EN contraindicated Persistent high GRVs If EN contraindicated
Anthropometric Weight, height Weight Weight Weight Weight, length, head Weight
monitoring circumference,
abdominal girth
Method of estimating Schofield equation White equation NS Estimated average NS WHO equation
energy requirements requirements
(Department of Health
1991)
Type of formula Nutrison Pediatric Kindercal or Pediasure Either Pedialyte (if never NS Start with Pedialyte and Commercially available
(<10 years) fed previously) or progress to EBM or ready-made solutions
Standard formulas EBM/formula formula (EBM or based on age
(>10 years) elemental formula if
heme+ stools)
Dose at initiation Started as 50% predicted 10 ml/hr 1 ml/hr for 4 hrs (<2.5 kg), Full target feeds initiated 3 ml/hr Start at 25% of target
PBMR on D1 or 25 ml/hr for 4 hrs at 2 hourly boluses volume
(>2.5 kg)
Target time of initiation <12 hr <6 hr NS <6 hr NS <24 hr
from ICU admission
Dose advancement 100% PBMR at D2 Increase by 5 ml/hr every Increase by 1 ml/hr Full target feeds Increase by 3 ml/hr every Increase by 25% every 4
125% PBMR at D3 2 hr (<2.5 kg) or 25 ml/hr progressed to 3 hourly 4 hr if tolerated hours if tolerated
150% PBMR at D4 (>2.5 kg) every 4 hr if or 4 hourly boluses
and D5 tolerated
GRV monitoring Checked GRV every 4 No routine measurement Checked GRV every 4 hr Checked GRV every 4 hr Checked GRV every 4 hr Checked GRV every 4 hr
hr (<25% of the feed of GRV once feeding (<50% of feed volume (<5 ml/kg or <300 ml (<50% of feed volume (<50% of feed volume
volume given accepted) started given accepted) is accepted, whichever given accepted) given was accepted)
lower)
Continuous or bolus/ All started with All started with All started with Continuous feeds All started with All started with continuous
interval feeding continuous feeds continuous feeds continuous feeds considered if continuous feeds feeds

Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014

persistently high GRV
Motility agents Cisapride started if there Metochlorpromide Metochlorpromide started Erythromycin started if Metochlopromide started NS
were high GRVs started if patient is if there were high GRVs there were high GRVsa if there were high GRVs
on a narcotic drip or
constipated
Acid suppressants NS NS Ranitidine if NBM NS Ranitidine or lansoprazole NS
for all patients
Gastric or small bowel All had intragastric NS Small bowel feeding if Small bowel feeding if NS NS
feeding feeding patients not tolerating persistent high GRVsa
intragastric feeding
Signs of intolerance NS Clearly stated Clearly stated Clearly stated Clearly stated NS
Nurse driven No Yes NS Yes NS Yes

Only 6 of 9 studies had description of the protocol in their published study. D1-5, day 1-5; EBM, expressed breast milk; EN, enteral nutrition; GRV, gastric residual volume; ICU, intensive care unit;
NBM, nil by mouth; NS, not specified; PBMR, predicted basal metabolic rate; PN, parenteral nutrition; WHO, World Health Organization.
a

37
These steps had to be discussed with the attending physician.
38 Journal of Parenteral and Enteral Nutrition 38(1)
Acknowledgments
We thank Peggy Fong Bih Yuh, librarian of KK Womens and
Childrens Hospital, for instruction on the search strategy. We
also appreciate the support of Duke-NUS/SingHealth Academic
Medicine Research Institute and the medical editing assistance of
Taara Madhavan (associate, Duke-NUS Graduate Medical
School).
References
1. Pollack MM, Wiley JS, Kanter R, Holbrook PR. Malnutrition in criti-
cally ill infants and children. JPEN J Parenter Enteral Nutr. 1982 Jan-
Feb;6(1):20-24.
2. Pollack MM, Ruttimann UE, Wiley JS. Nutritional depletions in critically
ill children: associations with physiologic instability and increased quan-
tity of care. JPEN J Parenter Enteral Nutr. 1985 May-Jun;9(3):309-313.
3. Hulst J, Joosten K, Zimmermann L, et al. Malnutrition in critically ill
children: from admission to 6 months after discharge. Clin Nutr. 2004
Apr;23(2):223-232.
4. Delgado AF, Okay TS, Leone C, Nichols B, Del Negro GM, Vaz FA.
Hospital malnutrition and inflammatory response in critically ill children
and adolescents admitted to a tertiary intensive care unit. Clinics. 2008
Jun;63(3):357-362.
5. Leite HP, Isatugo MK, Sawaki L, Fisberg M. Anthropometric nutritional
assessment of critically ill hospitalized children. Rev Paul Med. 1993 Jan-
Feb;111(1):309-313.
6. Mehta NM. Approach to enteral feeding in the PICU. Nutr Clin Pract.
2009 Jun-Jul;24(3):377-387.
7. McClave SA, Heyland DK. The physiologic response and associated
clinical benefits from provision of early enteral nutrition. Nutr Clin Pract.
2009 Jun-Jul;24(3):305-315.
8. Prieto MB, Cid JL-H. Malnutrition in the critically ill child: the impor-
tance of enteral nutrition. Int J Environ Res Pub Health. 2011;8(11):
4353-4366.
9. de Souza Menezes F, Leite HP, Koch Nogueira PC. Malnutrition as
an independent predictor of clinical outcome in critically ill children.
Nutrition. 2012 Mar;28(3):267-270.
10. Marik PE, Zaloga GP. Early enteral nutrition in acutely ill patients: a sys-
tematic review. Crit Care Med. 2001 Dec;29(12):2264-2270.
11. Chellis MJ, Sanders SV, Webster H, Dean JM, Jackson D. Early enteral
feeding in the pediatric intensive care unit. JPEN J Parenter Enteral
Nutr.1996 Jan-Feb;20(1):71-73.
12. Mehta NM, McAleer D, Hamilton S, et al. Challenges to optimal enteral
nutrition in a multidisciplinary pediatric intensive care unit. JPEN J
Parenter Enteral Nutr. 2010 Jan 1;34(1):38-45.
13. 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 Feb;27(1):65-71.
14. Joffe A, Anton N, Lequier L, et al. Nutritional support for critically ill
children. Cochrane Database Syst Rev. 2009;(2):CD005144.
15. Mehta NM, Bechard LJ, Cahill N, et al. Nutritional practices and their
relationship to clinical outcomes in critically ill childrenan international
multicenter cohort study. Crit Care Med. 2012 Jul;40(7):2204-2211.
16. Morgan J, Young L, McGuire W. Slow advancement of enteral feed vol-
umes to prevent necrotising enterocolitis in very low birth weight infants.
Cochrane Database Syst Rev. 2011;(3):CD001241.
17. Guyatt G, Oxman AD, Akl EA, et al. GRADE guidelines: 1.
IntroductionGRADE evidence profiles and summary of findings tables.
J Clin Epidemiol. 2011 Apr;64(4):383-394.
18. Schunemann HJ, Jaeschke R, Cook DJ, et al. An official ATS statement:
grading the quality of evidence and strength of recommendations in ATS
guidelines and recommendations. Am J Respir Crit Care Med. 2006 Sep
1;174(5):605-614.
Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014
19. Swiglo BA, Murad MH, Schunemann HJ, et al. A case for clarity, con-
sistency, and helpfulness: state-of-the-art clinical practice guidelines in
endocrinology using the Grading of Recommendations, Assessment,
Development, and Evaluation system. J Clin Endocrinol Metab. 2008
Mar;93(3):666-673.
20. Gottschlich MM, Jenkins ME, Mayes T, Khoury J, Kagan RJ, Warden
GD. The 2002 Clinical Research Award. An evaluation of the safety of
early vs delayed enteral support and effects on clinical, nutritional, and
endocrine outcomes after severe burns. J Burn Care Rehabil. 2002 Nov-
Dec;23(6):401-415.
21. del Castillo SL, McCulley ME, Khemani RG, et al. Reducing the inci-
dence of necrotizing enterocolitis in neonates with hypoplastic left heart
syndrome with the introduction of an enteral feed protocol. Pediatr Crit
Care Med. 2010 May;11(3):373-377.
22. Braudis NJ, Curley MA, Beaupre K, et al. Enteral feeding algorithm for
infants with hypoplastic left heart syndrome poststage I palliation. Pediatr
Crit Care Med. 2009 Jul;10(4):460-466.
23. Geukers VG, Dijsselhof ME, Sauerwein HP, Bos AP. Effect of a nurse-
driven feeding algorithm and the institution of a nutritional support team
on energy and macronutrient intake in critically ill children. e-SPEN J.
2012;7(1):0-eng.
24. Petrillo-Albarano T, Pettignano R, Asfaw M, Easley K. Use of a feeding
protocol to improve nutritional support through early, aggressive, enteral
nutrition in the pediatric intensive care unit. Pediatr Crit Care Med. 2006
Jul;7(4):340-344.
25. Meyer R, Harrison S, Sargent S, Ramnarayan P, Habibi P, Labadarios D.
The impact of enteral feeding protocols on nutritional support in critically
ill children. J Hum Nutr Diet. 2009 Oct;22(5):428-436.
26. Tume L, Latten L, Darbyshire A. An evaluation of enteral feeding practices
in critically ill children. Nurs Crit Care. 2010 Nov-Dec;15(6):291-299.
27. Briassoulis GC, Zavras NJ, Hatzis MT. Effectiveness and safety of a pro-
tocol for promotion of early intragastric feeding in critically ill children.
Pediatr Crit Care Med. 2001 Apr;2(2):113-121.
28. Ghanayem NS, Jaquiss RD, Cava JR, et al. Right ventricle-to-pulmonary
artery conduit versus Blalock-Taussig shunt: a hemodynamic comparison.
Ann Thorac Surg. 2006 Nov;82(5):1603-1609; discussion 9-10.
29. Raja SG, Atamanyuk I, Kostolny M, Tsang V. In hypoplastic left heart
patients is Sano shunt compared with modified Blalock-Taussig shunt
associated with deleterious effects on ventricular performance? Interact
Cardiovasc Thorac Surg. 2010 Apr;10(4):620-623.
30. Woolf SH, Grol R, Hutchinson A, Eccles M, Grimshaw J. Clinical guide-
lines: potential benefits, limitations, and harms of clinical guidelines.
BMJ. 1999 Feb 20;318(7182):527-530.
31. Helwig A, Bower D, Wolff M, Guse C. Residents find clinical practice
guidelines valuable as educational and clinical tools. Fam Med. 1998
Jun;30(6):431-435.
32. Restrepo RD, Fortenberry JD, Spainhour C, Stockwell J, Goodfellow LT.
Protocol-driven ventilator management in children: comparison to non-
protocol care. J Intensive Care Med. 2004 Sep-Oct;19(5):274-284.
33. Verhoeven JJ, Brand JB, van de Polder MM, Joosten KF. Management
of hyperglycemia in the pediatric intensive care unit; implementation
of a glucose control protocol. Pediatr Crit Care Med. 2009 Nov;10(6):
648-652.
34. Deeter KH, King MA, Ridling D, Irby GL, Lynn AM, Zimmerman
JJ. Successful implementation of a pediatric sedation protocol for
mechanically ventilated patients. Crit Care Med. 2011 Apr;39(4):
683-688.
35. Heyland DK, Cahill NE, Dhaliwal R, Sun X, Day AG, McClave SA.
Impact of enteral feeding protocols on enteral nutrition delivery: results
of a multicenter observational study. JPEN J Parenter Enteral Nutr. 2010
Nov-Dec;34(6):675-684.
36. Martin CM, Doig GS, Heyland DK, Morrison T, Sibbald WJ. Multicentre,
cluster-randomized clinical trial of algorithms for critical-care enteral and
parenteral therapy (ACCEPT). CMAJ. 2004 Jan 20;170(2):197-204.
Wong et al 39
37. Barr J, Hecht M, Flavin KE, Khorana A, Gould MK. Outcomes in criti-
cally ill patients before and after the implementation of an evidence-
based nutritional management protocol. Chest. 2004 Apr;125(4):
1446-1457.
38. Doig GS, Simpson F, Finfer S, et al. Effect of evidence-based feeding
guidelines on mortality of critically ill adults: a cluster randomized con-
trolled trial. JAMA. 2008 Dec 17;300(23):2731-2741.
39. Woien H, Bjork IT. Nutrition of the critically ill patient and effects of
implementing a nutritional support algorithm in ICU. J Clin Nurs. 2006
Feb;15(2):168-177.
40. Griffiths RD. Guidelines for nutrition in the critically ill: are we alto-
gether or in-the-altogether? JPEN J Parenter Enteral Nutr. 2010 Nov
1;34(6):595-597.
41. Mackenzie SL, Zygun DA, Whitmore BL, Doig CJ, Hameed SM.
Implementation of a nutrition support protocol increases the proportion
of mechanically ventilated patients reaching enteral nutrition targets in
the adult intensive care unit. JPEN J Parenter Enteral Nutr. 2005 Mar-
Apr;29(2):74-80.
Downloaded from pen.sagepub.com at University of Texas at El Paso on August 15, 2014
42. Blue JW, Colburn WA. Efficacy measures: surrogates or clinical out-
comes? J Clin Pharmacol. 1996 Sep;36(9):767-770.
43. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the
median, range, and the size of a sample. BMC Med Res Methodol. 2005;5:13.
44. Horn D, Chaboyer W. Gastric feeding in critically ill children: a random-
ized controlled trial. Am J Crit Care. 2003 Sep;12(5):461-468.
45. Horn D, Chaboyer W, Schluter PJ. Gastric residual volumes in critically
ill paediatric patients: a comparison of feeding regimens. Aust Crit Care.
2004 Aug;17(3):98-100, 2-3.
46. Pollack MM, Wiley JS, Holbrook PR. Early nutritional depletion in criti-
cally ill children. Crit Care Med. 1981 Aug;9(8):580-583.
47. Gramlich L, Kichian K, Pinilla J, Rodych NJ, Dhaliwal R, Heyland DK.
Does enteral nutrition compared to parenteral nutrition result in better out-
comes in critically ill adult patients? A systematic review of the literature.
Nutrition. 2004 Oct;20(10):843-848.
48. Dobson K, Scott A. Review of ICU nutrition support practices: imple-
menting the nurse-led enteral feeding algorithm. Nurs Crit Care. 2007
May-Jun;12(3):114-123.