552848

research-article2014
PENXXX10.1177/0148607114552848Journal of Parenteral and Enteral NutritionViswanathan et al

Original Communication
Journal of Parenteral and Enteral
Nutrition
A Standardized Slow Enteral Feeding Protocol and the Volume XX Number X
Month 201X 1­–11
Incidence of Necrotizing Enterocolitis in Extremely Low © 2014 American Society

Birth Weight Infants for Parenteral and Enteral Nutrition

DOI: 10.1177/0148607114552848
jpen.sagepub.com
hosted at
online.sagepub.com

Sreekanth Viswanathan, MD1; Kera McNelis, MD2; Dennis Super, MD, MPH3;
Douglas Einstadter, MD, MPH4; Sharon Groh-Wargo, PhD, RD, LD2;
and Marc Collin, MD2

Abstract
Background: Compared with early enteral feeds, the delayed introduction and slow advancement of enteral feedings to reduce the
incidence of necrotizing enterocolitis (NEC) are not well studied in extremely low birth weight (ELBW) infants. Objective: To study the
effects of a standardized slow enteral feeding (SSEF) protocol in ELBW infants. Methods: ELBW infants who followed an SSEF protocol
(September 2009 to December 2012) were compared with a similar group of historical controls (January 2003 to July 2009). Short-term
outcomes between the 2 groups were compared by propensity score (PS) analysis. Results: One hundred twenty-five infants in the SSEF
group were compared with 294 historical controls. Compared with the controls, feeding initiation day, full enteral feeding day, parenteral
nutrition (PN) days, and total central line days were longer in the SSEF group. There was no significant difference in overall NEC (5.6%
vs 11.2%, respectively; P = .10) or surgical NEC (1.6% vs 4.8%, respectively; P = .17) between the SSEF group and controls. However,
in infants with birth weight <750 g, NEC (2.1% vs 16.2%, respectively; P < .01) or combined NEC/death (12.8% vs 29.5%, respectively;
P = .03) was significantly less in the SSEF group compared with controls. In infants who survived to discharge, there was no significant
difference in the discharge weight or length of stay in PS-adjusted analysis. Conclusions: An SSEF protocol significantly reduces the
incidence of NEC and combined NEC/death in infants with birth weight <750 g. Despite taking longer to achieve full enteral feeding on
this protocol, surviving ELBW infants demonstrated comparable weight gain at discharge without prolonging their hospital stay. (JPEN
J Parenter Enteral Nutr. XXXX;XX:xx-xx.)

Keywords
necrotizing enterocolitis; preterm infants; extremely low birth weight infants; feeding protocol

Clinical Relevancy Statement Human Development (NICHD) research network outcome

data, the incidence of NEC was 11% for preterm infants born
Necrotizing enterocolitis (NEC) continues to be a significant at <28 weeks’ gestation.8 In addition to short-term complica-
cause of mortality and morbidity in extreme preterm infants. tions such as feeding intolerance, intestinal obstruction, and
The increased use of human breast milk and implementation of
standardized feeding protocols have helped to reduce the inci-
dence of NEC. In units where preterm formula is used when From the 1Division of Neonatology, Department of Pediatrics, Rainbow
breast milk is not available, this study has shown a way to sus- Babies and Children’s Hospital, Case Western Reserve University,
tain the reduction of NEC by slower standardized enteral feed- Cleveland, Ohio; 2Division of Neonatology, Department of Pediatrics,
MetroHealth Medical Center, Case Western Reserve University,
ing advancements, especially in babies with birth weight <750
Cleveland, Ohio; 3Division of Pediatric Critical Care, Department
g, which is the group with the highest risk of dying from NEC. of Pediatrics, MetroHealth Medical Center, Case Western Reserve
University, Cleveland, Ohio; and 4Department of Epidemiology and
Biostatistics, MetroHealth Medical Center, Case Western Reserve
Introduction University, Cleveland, Ohio.
Necrotizing enterocolitis (NEC) continues to be a devastating Financial disclosure: None declared.
neonatal illness, especially in extremely low birth weight
Received for publication July 24, 2014; accepted for publication August
(ELBW, birth weight [BW] ≤1000 g) preterm infants.1-3 In 5, 2014.
very low birth weight (VLBW, BW <1500 g) infants, the mean
Corresponding Author:
incidence of NEC ranges from 7%–9%, with an estimated case
Sreekanth Viswanathan, MD, Division of Neonatology, Department of
fatality rate of 15%–30% and the greatest mortality in infants Pediatrics, Rainbow Babies and Children’s Hospitals, Case Western
requiring surgery for NEC.1,4-7 In the most recently published Reserve University, 11100 Euclid Avenue, Cleveland, OH 44106, USA.
Eunice Kennedy Shriver National Institute of Child Health and Email: [email protected]

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015

2 Journal of Parenteral and Enteral Nutrition XX(X)

Table 1. Enteral Feeding Practices for ELBW Infants During the Respective Study Period.

Control Group SSEF Group

Characteristics Birth Weight <750 g Birth Weight = 750–1000 g Birth Weight <750 g Birth Weight = 750–1000 g
NPO days Not defined Not defined 14 7
Trophic feeding days 3–7 3–7 7 7
Trophic feeding volume 10 mL/kg/d 10 mL/kg/d 0.5 mL every 2 h 1 mL every 2 h
Feeding advancement 15 mL/kg/d 15–20 mL/kg/d 0.5 mL/feed every other day 0.5 mL/feed every day
Days to full feeds 18–22 16–20 44–52 32–36
(150 mL/kg/d)
Provide progressive feedings of human milk or In both weight groups, change 2 hourly to 3 hourly feeds
24-kcal/oz preterm formula every 2–3 h. Add when infant weight is >1250 g and then advance feeds
HMF when enteral feeds reach 150 mL/kg/d. by 1 mL/feed/d until full feeds. Add HMF (1:50 mL)
when enteral feeds reach 100 mL/kg/d and HMF
(1:25 mL) at 150 mL/kg/d.

ELBW, extremely low birth weight; HMF, human milk fortifier; NPO, nil per os; SSEF, standardized slow enteral feeding.

short gut syndrome, surviving infants, particularly infants with criteria.17,18 Due to the limited number of ELBW infants
surgical NEC, have poorer neurodevelopmental outcomes and enrolled in these studies, caution must be used when general-
represent a huge financial burden to the healthcare system.9 izing these results to all ELBW infants, the group at highest
The pathophysiology of NEC is poorly understood, but it is risk for developing NEC.16
likely a multifactorial disease.7,10 Immaturity of the intestinal The precise effect of enteral feeding advancement on the
tract, inappropriate responses to injuries, abnormal bacterial occurrence of NEC has not yet been thoroughly investigated in
colonization, and genetic predisposition have all been impli- ELBW infants. We hypothesized that exposure to a standard-
cated in the etiology of NEC.7 Since little progress has been ized slow enteral feeding (SSEF) protocol might reduce the
made in the management of NEC once it occurs, preventive incidence of NEC in ELBW infants without inducing signifi-
strategies are more likely to have a greater impact in reducing cant adverse events. To investigate this hypothesis, we carried
the mortality and morbidity from NEC.11 out a prospective study in a cohort of ELBW infants who fol-
Most preterm infants who develop NEC have received lowed the SSEF protocol and compared the short-term out-
enteral feeds. However, it remains unclear which aspects of comes with a historical control group of ELBW infants
feeding regimens affect the risk of NEC. Significant varia- admitted to the same neonatal intensive care unit (NICU) prior
tions in practice exist as to when feeds are initiated, how they to implementation of the SSEF protocol.
are advanced, and how feeding intolerance is managed in pre-
term infants.2 The modifiable risk factors related to enteral
feeding for the development of NEC in preterm infants include
Methods
the timing of introducing the feeds, the duration of trophic The study took place in the level III NICU at the MetroHealth
feeding and the rate of advancement of feeding, and the type Medical Center (MHMC). The NICU receives nearly 600
of milk (human milk vs formula feeding).12,13 Standardizing admissions per year, including approximately 50 ELBW infants,
the feeding regimen itself has been shown to reduce the inci- and serves a diverse, underserved inner city population in
dence of NEC.13 Cleveland, Ohio. The nonstandardized feeding guideline for
An immature enteric nervous system and intestinal dys- ELBW infants admitted to the MHMC NICU was replaced by
motility warrant gradual and cautious increments in enteral the SSEF protocol in August 2009 (see Table 1 for protocol
feedings. Observational studies have reported a higher inci- description). We stratified ELBW infants into 2 weight groups
dence of NEC in centers where enteral feeding is introduced (BW <750 g and BW 750–1000 g), and separate SSEF protocols
earlier and feeding volumes are advanced more quickly.12-15 were developed for each group. The SSEF protocols differed
Pietz et al15 reported a 0.4% incidence of NEC in 1158 VLBW from previous feeding guidelines by delaying the start of enteral
infants (~60% were ELBW infants) who followed a late-onset, feeding, with more days to prime the intestine and more cautious
slow, continuous drip feeding protocol. In contrast, a recent increments in feeding. In addition, powdered human milk forti-
meta-analysis of 5 randomized controlled trials (RCTs) com- fier (HMF) was introduced earlier during the SSEF protocol
paring slow advancement (<15–20 mL/kg/d) vs faster advance- when the enteral feed reached 100 mL/kg/d compared with 150
ment (30–35 mL/kg/d) did not detect any significant difference mL/kg/d with the previous guideline. At the attending physi-
in NEC or all-cause mortality.16 However, ELBW infants were cian’s discretion, the initiation of feeds was allowed to be
only included in 2 of the studies, and both had broad exclusion delayed, if indicated, but no infant was fed sooner than the

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015

Viswanathan et al 3

designated times specified by the protocol. Feeding intolerance cerebrospinal fluid (CSF) culture–positive sepsis/meningitis,
was defined as gastric residuals >2 mL for infants with BW cholestasis (direct bilirubin levels), metabolic bone disease
<750 g and 3 mL for infants between BW 750–1000 g or >50% (serum alkaline phosphatase [ALP] levels), patent ductus arte-
of the prior feeding, bile- or blood-stained aspirates, abdominal riosus (PDA), retinopathy of prematurity (ROP), and use of
distention/tenderness, or presence of blood in the stool. Feeding medications (postnatal steroids, inotropes, antibiotics, ibupro-
intolerance was quantified by the number of days that feeding fen/indomethacin). Late-onset sepsis/meningitis was defined
was withheld ≥24 hours. Parenteral nutrition (PN) with a mini- as clinical signs and symptoms consistent with sepsis occur-
mum 1 g/kg/d of protein was initiated on admission to the NICU ring >3 days after birth associated with the isolation of a caus-
during the whole study period. The protein content of starter PN ative organism from at least 1 blood or CSF culture. Patients’
was increased to 2.5 g/kg/d halfway through the SSEF study outcomes including mortality rate, length of hospital stay, and
period. Intravenous fat emulsion was discontinued when enteral weight at discharge from the NICU (among survivors) were
feedings reached 100 mL/kg/d, and PN was discontinued when recorded. The study was approved by the institutional review
enteral feedings reached 120 mL/kg/d. Human milk feeding was board of the MHMC.
encouraged, and if not available, standard preterm formula (24
cal/oz) was used. Donor breast milk (DBM) or probiotic prepa-
rations were not used during the entire study period.
Statistical Analysis
The incidence of NEC in ELBW infants of ≤30 weeks’ birth Based on the historical data, the incidence of NEC in ELBW
gestation admitted to the MHMC NICU between September infants was 11% at the study initiation (MHMC NICU statis-
2009 and December 2012, who followed the SSEF protocol tics). We determined that a prospective group sample size of
(SSEF group, prospective cohort), was compared with ELBW 125 ELBW infants compared with 300 controls would have
infants of ≤30 weeks’ gestation admitted to the MHMC NICU 80% power to detect a decrease in the incidence of NEC to
between January 2003 and July 2009, who followed a nonstan- 3.3%, using a 2-tailed 95% confidence interval (CI). This low
dardized feeding guideline (historical controls). We excluded value for the incidence of NEC was chosen based on the inci-
infants with major anomalies/known gastrointestinal anoma- dence of NEC observed in prior studies using comparable
lies, operative diagnosis of spontaneous intestinal perforation, late-onset slow enteral feeding.15 We performed an appropri-
as well as infants who developed NEC/died before the initia- ate bivariate analysis to identify the unadjusted differences
tion of feeds or were transferred to another facility before between the SSEF group and historical controls. All quantita-
reaching full enteral feeding. tive data are expressed as the mean ± standard deviation or
The primary outcome of the study was the incidence of median (interquartile range). A P value <.05 was considered
NEC in ELBW infants. We defined the occurrence of NEC as to be statistically significant. We also performed a propensity
Bell stage 2 or greater as per the modified Bell classification19 score (PS) analysis to calculate the adjusted SSEF group
or when diagnosed at surgery . Secondary outcomes included effect compared with controls on various outcomes. The PS
the incidence of NEC or death combined, discharge weight, analysis represents an improvement over traditional model-
late-onset sepsis, cholestasis, and metabolic bone disease of ing strategies. With PS methods, infants in the SSEF group
prematurity. We reviewed all medical records for infants’ are matched on a range of potentially confounding factors to
demographics, including gestational age (GA), BW, small for infants in the control group. The 2 groups can then be consid-
GA (SGA), sex, race, mode of delivery, Apgar scores at 1 and ered equivalent to each other if no statistical difference exists
5 minutes, exposure to antenatal steroids and chorioamnionitis, between the groups on all the covariates included in the
and severity of illness measured by the Score for Neonatal model. The 2 PS-matched groups are then compared with
Acute Physiology–Perinatal Extension (SNAP-PE). The each other on various outcomes; thus, the PS analysis repre-
SNAP-PE score is a 9-item neonatal illness severity and mor- sents a quasirandomized controlled design using observa-
tality risk score.20 It is calculated from data collected on the tional data. Using multiple logistic regression including
day of admission to the NICU, with points given for physiolog- baseline demographic and nutrition variables, we calculated a
ical items, BW, low Apgar score, and SGA. Nutrition data col- PS for entering the SSEF group for each infant. Analyses
lected include enteral feeding initiation day, trophic feeding using 1:1 greedy PS matching stratified by BW group as well
days, trophic feeding volume, days to reach full enteral feeding as PS weighting by the inverse PS to calculate the mean inter-
(150 mL/kg/d), human milk use, PN days, and the duration of vention effect were performed. Conditional logistic regres-
central line use (umbilical arterial line [UAC], umbilical sion/paired t test (PS matching) and survey design (PS
venous line, peripherally inserted central catheter, Broviac weighting) were performed based on the PS-matched/
lines, and total central line days [excluding UAC days]). We weighted pairs as well as direct PS-adjusted comparisons of
also reviewed the medical records for the number of days of the SSEF group to the controls on NEC and other secondary
mechanical ventilation, the incidence of chronic lung disease outcomes. Statistical software R version R-2.14.1 (R
(CLD; oxygen requirement at 36 weeks of corrected gesta- Foundation for Statistical Computing, Vienna, Austria) was
tion), intraventricular hemorrhaging (IVH), blood or used for the statistical analysis of the data.

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015

4 Journal of Parenteral and Enteral Nutrition XX(X)

Table 2. Demographics of ELBW Infants in the Control and SSEF Groups.

Characteristics Control Group (n = 294) SSEF Group (n = 125) P Value

Birth weight, g 754.1 ± 147.7 766.8 ± 155.5 .44
Birth weight percentile 34.4 ± 25.1 38.6 ± 25.6 .12
Birth gestation, wk 25.9 ± 1.8 25.7 ± 1.8 .33
Male sex, % 48.9 60.8 .03
Black race, % 60.5 57.6 .59
Small for gestational age, % 20.7 25.6 .30
1-minute Apgar score, median (IQR) 4 (3–6) 4 (2–6) .05
5-minute Apgar score, median (IQR) 7 (6–8) 7 (6–8) .10
Cesarean delivery, % 67.8 69.6 .74
Antenatal steroids, % 75.8 75.2 .99
Chorioamnionitis, % 17.3 14.4 .56
SNAP-PE 44.3 ± 14.7 40.9 ± 16.4 .05

Values are expressed as mean ± standard deviation unless otherwise indicated. ELBW, extremely low birth weight; IQR, interquartile range; SNAP-PE,
Score for Neonatal Acute Physiology–Perinatal Extension; SSEF, standardized slow enteral feeding.

Table 3. Nutrition Characteristics of ELBW Infants in the Control and SSEF Groups.

Characteristics Control Group (n = 294) SSEF Group (n = 125) P Value

Enteral feeding initiation day 11.1 ± 7.1 14.2 ± 7.3 <.001
Trophic feeding days 5.3 ± 3.2 7.9 ± 1.6 <.01
NPO days (feed start to full feeds) 6.0 ± 9.4 5.2 ± 6.6 .368
Enteral full feed day 35.1 ± 19.3 63.6 ± 19.1 <.001
Any human milk use, % 68.1 80.2 .013
Human milk to formula before full feeds, % 36.8 53.6 .001
Human milk on full feeds, % 38.3 28.5 .073
Regain birth weight day 15.9 ± 6.3 17.1 ± 6.5 .095
Weight on full feed, g 1046.5 ± 345.6 1665.9 ± 413.1 <.001
PN days 36.3 ± 22.5 62.7 ± 22.9 <.001

Values are expressed as mean ± standard deviation unless otherwise indicated. ELBW, extremely low birth weight; NPO, nil per os; PN, parenteral
nutrition; SSEF, standardized slow enteral feeding.

Results initiation of enteral feeds, took more days to reach full enteral
feeds, and required more PN days (Table 3). The rate of human
During the control study period, 391 ELBW infants were admit- milk initiation was greater in the SSEF group, but in both
ted to the MHMC NICU. A total of 97 infants met the exclusion groups, the majority of infants were on formula when they
criteria (72 infants died and 3 developed NEC before the initia- reached full enteral feeds (Table 3).
tion of enteral feeds, 3 infants were diagnosed with spontaneous The overall incidence of NEC (5.6% vs 11.2%, respec-
intestinal perforation, 13 infants were ≥31 weeks’ birth gestation, tively) and surgical NEC (1.6% vs 4.8%, respectively) were
5 infants were transferred out before reaching full enteral feeds, not significantly different between the SSEF group and con-
and 1 infant had a major congenital anomaly). During the SSEF trols (Table 4). However, in infants with BW <750 g, there was
study period, 145 ELBW infants were admitted to the MHMC a significant reduction in NEC (2.1% vs 16.2%, respectively; P
NICU, and 20 of them met the exclusion criteria (17 infants died < .01) in the SSEF group compared with controls (Table 4 and
before the initiation of enteral feeds, and 3 infants were ≥31 Figure 1). No infants with BW <750 g in the SSEF group
weeks’ birth gestation). The final study sample consisted of 294 developed surgical NEC compared with 7.8% in controls
infants in the control group and 125 infants in the SSEF group. (Figure 1). The timing of NEC onset was significantly delayed
The demographics of patients in the control and SSEF in the SSEF group compared with controls (57.9 ± 23.7 days
groups are shown in Table 2. Overall, there were no significant [range, 28–97 days] vs 31.2 ± 14.9 days [range, 9–66 days],
differences between the SSEF group and controls, except there respectively; P = .02] . The incidence of NEC before reaching
were more male infants in the SSEF group (Table 2). Compared full enteral feeds was similar between the SSEF group and con-
with controls, the SSEF group had significantly delayed trols (57.1% vs 54.5%, respectively; P > .99).

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015

Viswanathan et al 5

Table 4. Incidence of NEC in ELBW Infants in the Control and SSEF Groups.

NEC Incidence Control Group, n (%) SSEF Group, n (%) Odds Ratio (95% CI) P Value
All ELBW infants
Overall NEC ≥ stage 2 33/294 (11.2) 7/125 (5.6) 0.47 (0.17–1.12) .10
Surgical NEC 14/294 (4.8) 2/125 (1.6) 0.33 (0.04–1.45) .17
Birth weight <750 g
Overall NEC ≥ stage 2 21/129 (16.2) 1/47 (2.1) 0.11 (0.002–0.74) <.01
Surgical NEC 10/129 (7.8) 0/47 (0.0) 0.00 (0.000–1.18) .06
Birth weight = 750–1000 g
Overall NEC ≥ stage 2 12/165 (7.3) 6/78 (7.7) 1.06 (0.31–3.20) .99
Surgical NEC 4/165 (2.4) 2/78 (2.6) 1.06 (0.09–7.57) .99

CI, confidence interval; ELBW, extremely low birth weight; NEC, necrotizing enterocolitis; SSEF, standardized slow enteral feeding.

Figure 1. Incidence of necrotizing enterocolitis (NEC), surgical NEC, and NEC/death in infants with birth weight <750 g in the
standardized slow enteral feeding (SSEF) group compared with the control group. *P < .01. **P = .03.

The SSEF group required significantly more total central SSEF group had less CLD, hypotension requiring inotropic
line days compared with controls (60.1 ± 29 days vs 34.4 ± 31 medications, PDA ligation, and ROP laser surgery compared
days, respectively; P < .001) during their NICU stay (Table 5). with the controls (Table 5).
The incidence of culture-positive sepsis was not significantly In infants who survived to NICU discharge, the SSEF group
different between the 2 groups, but infants in the SSEF group had significantly greater body weight (2981 ± 912 g vs 2694 ±
developed infections later in their NICU stay and were exposed 842 g, respectively), with a similar length of NICU stay (110.2 ±
to a longer duration of antibiotics (Table 5). None of the infants 41 days vs 106.7 ± 43 days, respectively), compared with the
in the SSEF group with NEC had concomitant sepsis compared controls (Table 6). Extrauterine growth restriction (<10th per-
with 33% in controls (P = .16). The SSEF group had signifi- centile of weight for corrected GA) was significantly less in the
cantly higher peak ALP levels with no difference in the inci- SSEF group compared with the controls, while the percentage of
dence of cholestasis compared with controls (Table 5). The infants with a head circumference in <10th percentile at NICU

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015

6 Journal of Parenteral and Enteral Nutrition XX(X)

Table 5. Comorbidities Observed in ELBW Infants in the Control and SSEF Groups.

Characteristics Control Group SSEF Group P Value

Any sepsis, % 44.2 42.4 .75
Late-onset sepsis, % 42.9 40.8 .75
CONS sepsis, % 34.0 31.2 .65
Sepsis day of life 26.4 ± 27.8 38.2 ± 27.4 .01
Total antibiotic days 21.3 ± 20.3 25.4 ± 16.3 .03
Highest direct bilirubin 1.8 ± 2.6 2.1 ± 2.9 .10
Direct bilirubin ≥2 mg/dL, % 22.1 28.8 .13
Peak ALP, IU/dL 477.1 ± 211.6 545.9 ± 261.5 .01
UAC days 7.7 ± 4.5 7.5 ± 4.8 .70
UVC days 8.9 ± 4.5 6.6 ± 4.9 <.001
PICC line days 16.2 ± 16.9 51.4 ± 27.8 <.001
Broviac line days 9.8 ± 27.3 2.4 ± 9.9 <.001
Total central line daysa 34.4 ± 31.3 60.1 ± 29.5 <.001
Chronic lung disease, % 67.2 51.6 <.01
Mechanical ventilation days 39.3 ± 28.5 33.6 ± 31.1 .08
Postnatal steroid use, % 20.4 16.8 .42
Hypotension (inotrope use), % 46.6 27.2 <.001
Medical PDA, % 31.3 44.0 .01
Surgical PDA (ligation), % 42.5 27.2 <.01
Any IVH, % 37.0 34.4 .65
Grade 3/4 IVH, % 14.3 12.0 .64
Any ROP, % 82.3 73.6 .05
ROP laser treatment, % 28.9 18.4 .03

Values are expressed as mean ± standard deviation unless otherwise indicated. ALP, alkaline phosphatase; CONS, coagulase negative staphylococcal;
ELBW, extremely low birth weight; IVH, intraventricular hemorrhaging; PDA, patent ductus arteriosus; PICC, peripherally inserted central catheter;
ROP, retinopathy of prematurity; SSEF, standardized slow enteral feeding; UAC, umbilical arterial line; UVC, umbilical venous line.
a
Excluding UAC days.

Table 6. NICU Discharge Outcomes of ELBW Infants in the Control and SSEF Groups.

Outcome Control Group SSEF Group Odds Ratio (95% CI) P Value
Weight (median),a g 2694.2 ± 842.5 (2484) 2981.2 ± 912.0 (2885) <.01
Weight percentilea 8.6 ± 11.9 14.8 ± 15.1 <.001
Weight <10th percentile,a % 75.4 57.1 <.001
Head circumference <10th percentile,a % 48.4 38.1 .10
Length of stay (median),a d 106.7 ± 43.5 (99) 110.2 ± 41.5 (104) .47
Death, n (%) 37/294 (12.6) 6/125 (4.8) 0.35 (0.11–0.87) .02
<750 g 27/129 (20.9) 5/47 (10.6) 0.45 (0.16–1.25) .13
750–1000 g 10/165 (6.0) 1/78 (1.2) 0.20 (0.03–1.60) .11
Combined NEC/death, n (%) 57 (19.4) 13 (10.4) 0.48 (0.23–0.94) .03
<750 g 38/129 (29.5) 6/47 (12.8) 0.35 (0.14–0.89) .03
750–1000 g 19/165 (11.5) 7/78 (9.0) 0.76 (0.30–1.89) .65

Values are expressed as mean ± standard deviation unless otherwise indicated. CI, confidence interval; ELBW, extremely low birth weight; NEC,
necrotizing enterocolitis; NICU, neonatal intensive care unit; SSEF, standardized slow enteral feeding.
a
Only infants who survived to NICU discharge are included.

discharge was similar between the groups (Table 6). The SSEF combined NEC/death observed in the SSEF group was present
group had significantly lower death (4.8% vs 12.6%, respec- only in infants with BW <750 g (12.8% vs 29.5%, respectively;
tively; P = .02) and combined NEC/death (10.4% vs 19.4%, P = .03), while it was similar in the group with BW of 750–1000
respectively; P = .03) rates compared with controls (Table 6). g (9.0% vs 11.5%, respectively; P = .66). All the infants in the
However, in the subgroup analysis, the significant reduction in SSEF group with NEC survived to NICU discharge, while

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015

Viswanathan et al 7

Table 7. PS-Adjusted Outcomes in the SSEF Group Compared With the Control Group.

Outcome Unadjusted 1:1 PS Matching PS Weighting Direct PS Adjustment

NEC 0.47 (0.17 to 1.12) 0.44 (0.17 to 1.11) 0.45 (0.19 to 1.10) 0.41 (0.15 to 1.01)
Death 0.35 (0.11 to 0.87) 0.30 (0.11 to 0.78) 0.27 (0.11 to 0.68) 0.28 (0.10 to 0.66)
Combined NEC/death 0.48 (0.23 to 0.94) 0.42 (0.21 to 0.86) 0.42 (0.22 to 0.83) 0.40 (0.19 to 0.80)
Late-onset sepsis 0.92 (0.60 to 1.41) 0.85 (0.51 to 1.40) 0.87 (0.55 to 1.36) 0.84 (0.50 to 1.39)
Cholestasisa 1.47 (0.92 to 2.38) 1.12 (0.64 to 1.95) 1.27 (0.76 to 2.13) 1.08 (0.62 to 1.90)
Peak ALP, IU/dL 68.8 (20.9 to 116.9) 64.0 (3.6 to 124.3) 68.1 (12.6 to 123.6) 61.2 (0.62 to 121.8)
Discharge weight,b g 287.1 (98.6 to 475.6) 132.0 (–144.6 to 408.7) 202.0 (–18.0 to 422.0) 115.3 (–133.8 to 364.4)
Length of stay,b d 3.50 (–5.9 to 12.8) 6.12 (–4.1 to 16.4) 0.89 (–9.0 to 10.7) 5.7 (–4.8 to 16.1)

Values are expressed as odds ratio (95% confidence interval) for proportions for NEC, death, combined NEC/death, late-onset sepsis, and cholestasis
or estimated mean difference (95% confidence interval) for continuous variables for peak ALP, discharge weight, and length of stay . ALP, alkaline
phosphatase; NEC, necrotizing enterocolitis; PS, propensity score; SSEF, standardized slow enteral feeding.
a
Direct bilirubin ≥2 mg/dL.
b
Only infants who survived to NICU discharge are included.

39.4% (13/33) (<750 g: 10/21 [47.6%]; 750–1000 g: 3/12 potentially associated with death, an additional logistic regres-
[25%]) of infants with NEC in the control group died (P = .07). sion was performed to determine the adjusted SSEF group
NEC was an attributable cause of death in 35.1% (13/37) (<750 effect on the combined NEC/death outcome using all the
g: 10/27 [37.0%]; 750–1000 g: 3/10 [30%]) of the controls com- covariates potentially related to combined NEC/death with a P
pared with 0% (0/6) in the SSEF group. value ≤.10 in the unadjusted analysis (total antibiotics days,
To calculate the PS, the following baseline (BW, birth ges- cholestasis, peak ALP, total central line days, CLD, inotropic
tation, sex, race, SGA, mode of delivery, antenatal steroid use, PDA ligation, ROP laser). When controlling for these fac-
exposure, pre-eclampsia, maternal diabetes, maternal smoking, tors, being in the SSEF group was independently associated
substance abuse, magnesium tocolysis, chorioamnionitis, with a significant reduction in combined NEC/death (OR,
SNAP-PE score, Apgar scores of 1 and 5 minutes, early sepsis, 0.30; 95% CI, 0.12–0.71; P < .01).
UAC days, IVH) and nutrition (human milk as initial milk)
variables were included in the multiple logistic regression
model. The SSEF group had a mean PS of 0.36 ± 0.12, whereas
Discussion
the control group had a mean PS of 0.27 ± 0.15 (P < .001). In Our study demonstrates that an SSEF protocol with maternal
both the PS-matched groups (1:1 match of all infants in the human milk/preterm formula reduces the incidence of NEC
SSEF group to the nearest PS-matched controls without and combined NEC/death in infants with BW <750 g. These
replacement) and the PS-weighted groups, all the covariates results are comparable to those in studies that used an exclu-
included in the model were well balanced (none with a P value sive human milk diet for reducing NEC. However, due to the
<.05). The important PS-adjusted outcomes using various PS prolonged nature of the SSEF protocol, these infants were
methods are summarized in Table 7. The incidence of NEC exposed to longer central line and PN days. Despite taking lon-
was not significantly different between the SSEF group and ger to achieve full enteral feeding with the SSEF protocol,
controls in PS score–adjusted analysis. The PS-adjusted all- these infants demonstrated comparable weight gain at NICU
cause death and combined NEC/death rates were significantly discharge without prolonging their hospital stay.
lower in the SSEF group compared with controls. In the sub- Epidemiological studies have found that compared with non-
group analysis of infants with BW <750 g, the SSEF group had Hispanic white infants, non-Hispanic black infants have a higher
a significantly lower incidence of NEC (odds ratio [OR], 0.14; incidence of NEC.21 The incidence of NEC is also reported to be
95% CI, 0.007–0.74) and combined NEC/death (OR, 0.32; higher in male infants.22 Holman et al23 described the trends and
95% CI, 0.11–0.82) compared with controls , while both of risk factors for infant mortality in the United States and showed
these outcomes were similar in the 750–1000 g weight group. that death from NEC was highest in VLBW infants who were
The PS-adjusted incidence of late-onset sepsis and cholestasis black and male. In addition, the SNAP-PE score in the range of
were similar between the groups; however, the peak ALP level 40–49 observed in our study population represents a high neona-
was significantly higher in the SSEF group compared with the tal severity illness score, with a predicted mortality rate of
controls. In infants who survived to NICU discharge, the 15.9%.20 Thus, the demographic characteristics of ELBW
PS-adjusted discharge weight and length of stay were similar infants in our study (higher proportion of black male infants)
between the SSEF group and controls. represent a group at particularly high risk for developing NEC.
Since there were significant differences in comorbidities Observational data suggest that delaying the introduction of
between the SSEF and control groups (Table 5) that may be enteral feeds until 5–10 days postnatally reduces the risk of

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015

8 Journal of Parenteral and Enteral Nutrition XX(X)

NEC in VLBW infants.12,13,15 However, prolonged nil per os NEC in infants fed advancing volumes compared with those
(NPO) status may cause atrophy of the intestinal mucosa, fed with a minimal enteral feeding volume (10.4% vs 1.4%,
delayed development of absorptive function, decreased motil- respectively; P = .03) .30 Since it is difficult to blind the care-
ity, and development of proinflammatory changes.24 In various givers to the group assignment, measurement bias may have
observational studies, infants who received early trophic feeds overestimated the NEC rate in the advancing feed group.
were reported to have better feeding tolerance, improved However, the large difference in the NEC rate suggests that one
growth, reduced length of hospitalization, and decreased likeli- should be cautious with the advancement of aggressive feeding
hood of sepsis compared with infants who received delayed in the early days of enteral feeding, especially when using
enteral feeds.25-27 However, these potential short-term benefits formula.
of early enteral feeding were not shown to demonstrate a dif- In a recently reported multicenter RCT, the incidence of
ference in preventing NEC.16 A meta-analysis of 9 RCTs did NEC and surgical NEC in preterm infants with BW <1250 g,
not find a significant difference in the NEC rate, time to regain exclusively fed human milk, were 5.8% and 1.4% compared
BW, time to reach full enteral feeds, incidence of invasive with 15.9% and 10.1% in infants who were fed human milk
infections, duration of hospital stay, or all-cause mortality supplemented with bovine milk–based products, respectively.31
between early trophic feeding and enteral fasting (defined as The infants included in this study were less at risk for NEC
NPO for 5–7 days).16 The largest RCT that included ELBW (21.6% black infants, 9.6% SGA infants; mean BW, 925 g)
infants randomized appropriately grown infants of 26–30 compared with our study participants (57.6% black infants,
weeks’ gestation to enteral fasting or trophic feeding for the 25.6% SGA infants; mean BW, 766 g). The rates of feeding in
first 14 days of life.28 There was no difference in the NEC rate the Sullivan et al31 study (up to 5 days of trophic feeding [10–
(15.9% vs 11.2%, respectively) and days to reach full enteral 20 mL/kg/d], followed by 10–20 mL/kg/d of feed advance-
feeds (35 days vs 32 days, respectively) between the trophic ment) were also “slow” as per the definition set by Cochrane
feeding and enteral fasting groups.28 However, the primary Reviews (feed advancement <24 mL/kg/d). The days to full
outcome of the several included RCTs was not the incidence of enteral feeding and days of PN (22 and 21 days, respectively)
NEC but rather feeding intolerance or time to reach full enteral were significantly less compared with our study results (63 and
feeds. Additionally, only a minority of the participants in the 62 days, respectively). Subgroup data (BW <1000 g) from this
included trials were ELBW infants. In our study, the mean ini- study were not available to compare other outcomes such as
tial NPO period was only 3 days longer in the SSEF group sepsis, cholestasis, central line days, and hospital length of
compared with controls (13 days vs 10 days, respectively), and stay. With our SSEF protocol, we have achieved a similar rate
the time to reach full enteral feeds was significantly longer of NEC to infants fed on an exclusively human milk–based
(63.6 ± 19.1 days vs 35.1 ± 19.3 days, respectively) (Table 3). diet. In Sullivan et al’s31 study, even after following a standard-
This suggests that the slow feeding advancement is more likely ized feeding regimen, the high rate of NEC (15.9%) in infants
associated with NEC reduction rather than the initial NPO who received human milk supplemented with bovine milk–
period. based products suggests that a much slower advancement of
An immature enteric nervous system and intestinal dys- feeds such as in our SSEF protocol may be necessary to reduce
motility warrant gradual and cautious increments in enteral NEC in infants receiving bovine milk–based formula or
feedings. The advancement of aggressive enteral feeding is a fortifiers.
risk factor associated with NEC.29,30 However, a recent meta- Although no RCT has compared the effect of mother’s own
analysis of 5 RCTs between slow advancement (<15–20 mL/ milk (MM) vs formula on NEC or death, there is widespread
kg/d) vs faster advancement (30–35 mL/kg/d) did not detect consensus about the short- and long-term benefits of MM in
any significant difference in NEC or all-cause mortality.16 preterm infants including its protective effect on NEC.32,33 A
Infants who had slower advancement took longer to regain BW recently reported quality improvement project in California to
(difference of 2–6 days) and to reach full enteral feeds (differ- increase human milk use in VLBW infants resulted in a reduc-
ence of 2–5 days).16 However, ELBW infants were only tion of NEC from 7% to 2.4%.34 In another prospective cohort
included in 2 studies, and both had broad exclusion criteria.17,18 study of VLBW infants, infants who received ≥50% enteral
One RCT that considered NEC as the primary outcome evalu- feeds with human milk within the first 14 days of life had one
ated the effect of stable (20 mL/kg/d without advancement) vs sixth the odds of developing NEC compared with infants who
advancing (20 mL/kg/d to goal of 140 mL/kg/d) feeding vol- received <50% enteral feeds with human milk (3.2% vs 10.6%,
umes for a 10-day period in non-SGA VLBW infants.30 Enteral respectively; OR, 0.17; 95% CI, 0.04–0.68; P = .01).35 The
feeds were initiated at a mean age of 10 days, as feeds were higher rate of human milk initiation in the SSEF group com-
initiated only after the removal of umbilical catheters and the pared with controls could be a potential confounder in our
discontinuation of inotropic medications. Only one third of the study, but this difference was balanced by PS methods, and the
infants received human milk, and the remainder were fed outcomes were unchanged after PS analysis. If MM is not avail-
24-cal/oz preterm formula. This study was prematurely termi- able, the American Academy of Pediatrics (AAP) policy on
nated because they found a significantly higher incidence of breast feeding advocates DBM as suitable alternative feeding.33

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015

Viswanathan et al 9

Table 8. Key Discharge Outcome Variables of ELBW Infants for the SSEF Group and the VON.

Outcome SSEF Group VON,a 2010–2012

Mean weight,b g 2981.2 2816.1
Weight <10th percentile, % 57.1 55.0
Head circumference <10th percentile, % 38.1 39.0
Mean length of stay (median),c d 110.2 (104) 102.2 (99.3)

ELBW, extremely low birth weight; SSEF, standardized slow enteral feeding; VON, Vermont Oxford Network.
a
United States centers only.
b
Infants who survived to discharge “home.”
c
Total hospital stay by final disposition “alive.”

A meta-analysis of studies comparing DBM and formula dem- adoption of bundle strategies to prevent central line–associated
onstrated that preterm infants fed with formula had more than bloodstream infections potentially helped us to control the
twice the odds of NEC compared with infants fed with DBM.36,37 infection rate during the prospective study period. The statisti-
However, these studies included only a minority of extremely cally significantly higher peak ALP level in the SSEF group vs
preterm infants, and the effects of DBM combined with HMF the controls (545 IU/dL vs 477 IU/dL, respectively) reflects the
were not evaluated.31 An RCT that compared fortified, pasteur- potential nutrition-related harm associated with the SSEF pro-
ized DBM and preterm formula, both used as supplements tocol; however, the magnitude of difference may not be clini-
when MM was not available, did not find a protective effect of cally meaningful. We believe that the potential benefits of the
DBM on the combined incidence of late-onset sepsis and SSEF protocol outweigh the risks, as these infants had a sig-
NEC.32 Similarly, in Sullivan et al’s31 study, a high incidence of nificant reduction in NEC/death and demonstrated comparable
NEC (15.6%) was observed in infants who received human weight gain and head growth at NICU discharge without pro-
milk (MM or DBM) supplemented with bovine milk–based longing the NICU stay.
HMF. The use of MM/DBM may help in the earlier initiation In Sullivan et al’s31 study of infants ≤1250 g, the combined
and attainment of full enteral feeding, but slowing the feeding outcome of NEC/death was significantly less common in
advancement after supplementing with bovine milk–based exclusively human milk–fed infants compared with human
HMF may help the immature intestine to adapt and reduce the milk supplemented with a bovine milk–based product (7.3% vs
risk for developing NEC. Diet-dependent modification to the 20%, respectively). With the SSEF protocol, we have demon-
standardized feeding regimen may be needed for the sustained strated a similar reduction in NEC/death in ELBW infants
prevention of NEC, for example, slower advancement of feeds (10.4% vs 19.4%, respectively). This is even more significant,
in infants fed with bovine milk–based products alone or mixed as most of the reduction in NEC/death happened in infants with
with MM/DBM compared with an exclusive human milk diet. BW <750 g (12.8% vs 29.5%, respectively), which is the group
In a population-based cohort study of ELBW infants, NEC with a reported NEC mortality rate of 40%–60% as per the
occurred at a mean postnatal age of 32 days,38 similar to our largest NEC study cohort published.6 Based on our study
controls. The nearly doubled time to the onset of NEC in the results, the number of infants with BW <750 g that would be
SSEF group suggests that delaying the establishment of enteral needed to treat with the SSEF protocol to prevent 1 case of
feeds also delays the onset of NEC. This is also consistent with NEC is 7, and the number needed to treat to prevent 1 com-
the observation that NEC commonly occurs when the feeding bined NEC/death is 6.
volume exceeds 100 mL/kg/d.17,30,39 The need for surgical Key discharge outcome variables of the SSEF group com-
interventions is often used as a surrogate marker for the sever- pared with the corresponding period’s Vermont Oxford
ity of NEC.6,7,40 About 50% of infants <750 g who developed Network (VON) data for ELBW infants (2010–2012 in the
NEC in our control group progressed to surgical NEC com- United States only) are given in Table 8.41 The SSEF group has
pared with none in the SSEF group (Table 4). It is plausible comparable growth outcomes at NICU discharge to the VON
that by using an SSEF protocol, these extremely premature cohort. We believe that the modest increase in the length of
infants are more physiologically mature and have better pro- NICU stay is due to the improved survival of infants with BW
tective mechanisms to tolerate the delayed onset of NEC, lead- <750 g.
ing to a lower risk of progressing to advanced NEC that One of the strengths of our study is the use of NEC inci-
requires a surgical intervention. dence as the primary outcome. Most of the previous single-
Since conservative feeding strategies are associated with center studies were underpowered to detect a difference in the
the prolonged use of PN and central line days, they may alter NEC rate due to the relatively low incidence of NEC. However,
other competing outcomes, especially the rate of nosocomial our study is not without limitations. Only about one third of the
infections. There was no increase in the incidence of late-onset infants remained on human milk at full enteral feeds in spite of
sepsis in the SSEF group compared with controls; however, the the higher rate of human milk initiation (80.2%). Our results

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015

10 Journal of Parenteral and Enteral Nutrition XX(X)

are not isolated, as 1 study reported that only 30% of mothers 10. Henry MC, Moss RL. Necrotizing enterocolitis. Annu Rev Med.
were able to supply 100% of their extremely premature infants’ 2009;60:111-124.
11. Neu J, Mihatsch W. Recent developments in necrotizing enterocolitis.
needs.32 This suggests that our outcomes are not generalizable JPEN J Parenter Enteral Nutr. 2012;36(1 Suppl):30S-35S.
to the population fed mainly with a human milk diet. Since our 12. Henderson G, Craig S, Brocklehurst P, McGuire W. Enteral feeding regi-
study spans more than a decade, it is possible that practice mens and necrotising enterocolitis in preterm infants: a multicentre case-
changes in the care of ELBW infants have affected the out- control study. Arch Dis Child Fetal Neonatal Ed. 2009;94(2):F120-F123.
comes reported. For example, more medical treatments of PDA 13. Patole SK, de Klerk N. Impact of standardised feeding regimens on
incidence of neonatal necrotising enterocolitis: a systematic review and
compared with primary PDA ligation and earlier extubation meta-analysis of observational studies. Arch Dis Child Fetal Neonatal Ed.
and aggressive noninvasive ventilation in the SSEF group 2005;90:F147-F151.
reflect current neonatal practice. Also, we have reported only 14. Fanaroff AA, Stoll BJ, Wright LL, et al. Trends in neonatal morbidity
the short-term outcomes associated with the SSEF protocol, as and mortality for very low birth weight infants. Am J Obstet Gynecol.
the study was not designed to assess the long-term neurodevel- 2007;196:147.e1-147.e8.
15. Pietz J, Achanti B, Lilien L, et al. Prevention of necrotizing enteroco-
opmental outcomes. litis in preterm infants: a 20-year experience. Pediatrics. 2007;119:
e164-e170.
16. Morgan J, Young L, McGuire W. Slowly advancing milk feeds does not
Conclusion reduce the risk of necrotising enterocolitis in very low birth weight infants.
NEC is a devastating disease with high morbidity and mortal- Cochrane Database Syst Rev;28:CD001241.
17. Rayyis SF, Ambalavanan N, Wright L, Carlo WA. Randomized trial of
ity. Enteral feeding practices and the type of milk used repre- “slow” versus “fast” feed advancements on the incidence of necrotizing
sent 2 major modifiable risk factors of NEC in ELBW infants. enterocolitis in very low birth weight infants. J Pediatr. 1999;134(3):
Our study demonstrates that when using human milk or pre- 293-297.
term formula, following an SSEF protocol helps to reduce 18. Karagol BS, Zenciroglu A, Okumus N, Polin RA. Randomised controlled
NEC and combined NEC/death in infants with BW <750 g. We trial of slow versus rapid enteral feeding advancements on the clinical
outcomes of preterm infants with 750-1250g. JPEN J Parenter Enteral
believe that the benefits of reducing the combined NEC/death Nutr. 2013;37(2):223-228.
rate outweigh the potential nutrition-related harm associated 19. Walsh MC, Kliegman RM. Necrotizing enterocolitis: treatment based on
with the delayed initiation and slower advancement of enteral staging criteria. Pediatr Clin North Am. 1986;33(1):179-201.
feeds, especially when using formula as the initiation milk. 20. Richardson DK, Corcoran JD, Escobar GJ, Lee SK. SNAP-II and
However, because the etiology of NEC is multifactorial, only SNAPPE-II: simplified newborn illness severity and mortality risk scores.
J Pediatr. 2001;138(1):92-100.
adequately powered, vigorously conducted randomized trials 21. Llanos AR, Moss ME, Pinzon MC, Dye T, Sinkin RA, Kendig JW.
with sufficient follow-up can conclusively assess the effect of Epidemiology of neonatal necrotizing enterocolitis: a population-based
slow feeding advancement on NEC and other related short- study. Paediatr Perinat Epidemiol. 2002;16:342-349.
and long-term outcomes in ELBW infants. 22. Guner YS, Friedlich P, Wee CP, Dorey F, Camerini V, Upperman JS.
State-based analysis of necrotizing enterocolitis outcomes. J Surg Res.
2009;157(1):21-29.
References 23. Holman RC, Stoll BJ, Clarke MJ, Glass RI. The epidemiology of necrotiz-
1. Neu J, Walker WA. Necrotizing enterocolitis. N Engl J Med. ing enterocolitis infant mortality in the United States. Am J Public Health.
2011;364(3):255-264. 1997;87:2026-2031.
2. Athalye-Jape G, More K, Patole S. Progress in the field of necrotising 24. Kudsk KA. Current aspects of mucosal immunology and its influence by
enterocolitis: year 2012. J Matern Fetal Neonatal Med. 2013;26(7): nutrition. Am J Surg. 2002;183:390-398.
625-632. 25. De Silva A, Jones PW, Spencer SA. Does human milk reduce infec-
3. Holman RC, Stoll BJ, Curns AT, Yorita KL, Steiner CA, Schonberger LB. tion rates in preterm infants? A systematic review. Arch Dis Child Fetal
Necrotising enterocolitis hospitalisations among neonates in the United Neonatal Ed. 2004;89:F509-F513.
States. Paediatr Perinat Epidemiol. 2006;20(6):498-506. 26. Flidel-Rimon O, Branski D, Shinwell ES. The fear of necrotizing entero-
4. Fitzgibbons SC, Ching Y, Yu D, et al. Mortality of necrotizing entero- colitis versus achieving optimal growth in preterm infants: an opinion.
colitis expressed by birth weight categories. J Pediatr Surg. 2009;44(6): Acta Paediatr. 2006;95(11):1341-1344.
1072-1075, discussion 1075-1076. 27. McClure RJ. Trophic feeding of the preterm infant. Acta Paediatr Suppl.
5. Clark RH, Gordon P, Walker WM, et al. Characteristics of patients who 2001;90(436):19-21.
die of necrotizing enterocolitis. J Perinatol. 2012;32(3):199-204.8. 28. Schanler RJ, Shulman RJ, Lau C. Feeding strategies for premature infants:
6. Hull MA, Fisher JG, Gutierrez IM, et al. Mortality and management of beneficial outcomes of feeding fortified human milk vs. preterm formula.
surgical necrotizing enterocolitis in very low birth weight neonates: a pro- Pediatrics. 1999;103:1150-1157.
spective cohort study. J Am Coll Surg. 2014;218(6):1148-1155. 29. Anderson DM, Kliegman RM. The relationship of neonatal alimentation
7. Lin PW, Nasr TR, Stoll BJ. Necrotizing enterocolitis: recent scien- practices to the occurrence of endemic necrotizing enterocolitis. Am J
tific advances in pathophysiology and prevention. Semin Perinatol. Perinatol. 1991;8:62-67.
2008;32(2):70-82. 30. Berseth CL, Bisquera JA, Paje VU. Prolonging small feeding volumes
8. Stoll BJ, Hansen NI, Bell EF, et al. Neonatal outcomes of extremely pre- early in life decreases the incidence of necrotizing enterocolitis in very
term infants from the NICHD Neonatal Research Network. Pediatrics. low birth weight infants. Pediatrics. 2003;111:529-534.
2010;126:443-456. 31. Sullivan S, Schanler RJ, Kim JH, et al. An exclusively human milk-based
9. Bisquera JA, Cooper TR, Berseth CL. Impact of necrotizing enterocolitis diet is associated with a lower rate of necrotizing enterocolitis than a diet
on length of stay and hospital charges in very low birth weight infants. of human milk and bovine milk-based products. J Pediatr. 2010;156(4):
Pediatrics. 2002;109(3):423-428. 562-567.

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015

Viswanathan et al 11

32. Schanler RJ, Lau C, Hurst NM, Smith EO. Randomized trial of donor human 37. Boyd CA, Quigley MA, Brocklehurst P. Donor breast milk versus infant
milk versus preterm formula as substitutes for mothers’ own milk in the feed- formula for preterm infants: a systematic review and meta-analysis. Arch
ing of extremely premature infants. Pediatrics. 2005;116(2):400-406. Dis Child Fetal Neonatal Ed. 2007;92:F169-F175.
33. Eidelman AI, Schanler RJ. Section on Breastfeeding Executive 38. Yee WH, Soraisham AS, Shah VS, et al. Incidence and timing of pre-
Committee. Breastfeeding and the use of human milk: policy statement. sentation of necrotizing enterocolitis in preterm infants. Pediatrics.
Pediatrics. 2012;129:e827-e841. 2012;129(2):e298-e304.
34. Lee HC, Kurtin PS, Wight NE, et al. A quality improvement project to 39. Owens L, Berseth CL. Is there a volume threshold for enteral feeding and
increase breast milk use in very low birth weight infants. Pediatrics. necrotizing enterocolitis? Pediatr Res. 1995;37:315A.
2012;130:e1679-e1687. 40. Guthrie SO, Gordon PV, Thomas V, Thorp JA, Peabody J, Clark RH.
35. Sisk PM, Lovelady CA, Dillard RG, Gruber KJ, O’Shea TM. Early human Necrotizing enterocolitis among neonates in the United States. J Perinatol.
milk feedings associated with a lower risk of necrotizing enterocolitis in 2003;23:278-285.
very low birth weight infants. J Perinatol. 2007;27(7):428-433. 41. Vermont Oxford Network. Database of very low birth weight infants. 2010-
36. Quigley M, McGuire W. Formula versus donor breast milk for feeding preterm 2012. Nightingale Internet Reporting System. Available at: http://www.
or low birth weight infants. Cochrane Database Syst Rev. 2014;4:CD002971. vtoxford.org/. Accessed July 2, 2014.

Downloaded from pen.sagepub.com at GEORGIAN COURT UNIV on April 29, 2015