European Journal of Clinical Nutrition (2009) 63, 580584

& 2009 Macmillan Publishers Limited All rights reserved 0954-3007/09 $32.00
www.nature.com/ejcn

ORIGINAL ARTICLE

Does early enteral feeding of very low birth weight

infants increase the risk of necrotizing enterocolitis?
FC
akmak Celik1, C Aygun2 and E C
etinoglu3
1
Department of Pediatrics, Ondokuz Mays University Faculty of Medicine, Samsun, Turkey; 2Department of Neonatology, Ondokuz
Mays University Faculty of Medicine, Samsun, Turkey and 3Department of Public Health, Ondokuz Mays University Faculty of
Medicine, Samsun, Turkey

Background/Objective: In this retrospective study, we intended to test whether early enteral feeding (EEF) of very low birth
weight (VLBW) preterm babies increases the risk of necrotizing enterocolitis (NEC) or not.
Subjects and Methods: Overall, 297 VLBW preterm babies admitted to the neonatal intensive care unit (NICU) between April
2003 and April 2006 were included. The study consisted of two periods: the first period was between April 2003 and October
2004, when babies were not fed enterally until they were extubated (167 preterm VLBWs). The second period was between
November 2004 and April 2006, when babies were fed even when they were intubated, starting preferably on the first day of life
(130 preterm VLBWs). Criteria for withholding enteral feeding in both periods were hypotension necessitating vasopressor agent
use, abdominal distention, abdominal tenderness and suspected or proven NEC. Possible risk factors for NEC were also
recorded.
Results: The overall incidence of NEC in VLBW preterm babies was 6.7% and did not differ between the two study periods: 7.2%
in the late and 6.2% in the EEF regimens. On logistic regression analysis, the most important risk factors associated with NEC
were sepsis (Po0.001) and blood culture positivity (Po0.001). The average daily weight gain was significantly higher in the
early fed babies (P 0.011).
Conclusions: The EEF of VLBW preterm babies does not increase the risk of NEC. Increased daily weight gain is an important
reason to feed these babies earlier.

European Journal of Clinical Nutrition (2009) 63, 580584; doi:10.1038/sj.ejcn.1602957; published online 28 November 2007
Keywords: prematurity; necrotizing enterocolitis; very low birth weight; enteral feeding; septicemia; neonatal intensive care unit

Introduction
Necrotizing enterocolitis (NEC) is the most common and
important gastrointestinal disease of the newborn that needs
emergency intervention . The incidence of NEC varies across
different neonatal units and periods, ranging from 1 to 22%
in very low birth weight (VLBW) infants (Lin and Stoll,
2006). In the latest NICHD neonatal network cohort (1999
2001), about 7% of VLBW babies developed proven NEC
(4stage II), with about half undergoing surgery (Guillet
et al., 2006). In this cohort, NEC rates were inversely related
to birth weight. Intestinal immaturity due to prematurity,

Correspondence: Professor C Aygun, Department of Neonatology, Ondokuz

Mays University Faculty of Medicine, Samsun 55139, Turkey.
E-mail: cananaygun4@yahoo.com
Received 22 February 2007; revised 15 October 2007; accepted 16 October
2007; published online 28 November 2007

poor intestinal motility, umbilical catheter placement,

systemic infections, early enteral feeding (EEF) and rapid
increment of enteral feedings are cited as major risk factors
for NEC in addition to indomethacin and corticosteroid use
(McKeown et al., 1992; Lin and Stoll, 2006). The risk for NEC
postpones enteral feeding of these babies in many neonatal
intensive care units (NICUs). In addition, there are studies
emphasizing the importance of late enteral feedings in the
prevention of NEC (Pietz et al., 2007). As a result, in many
NICUs, VLBW babies are fed enterally when they are many
days old, after they are extubated and stabilized.
Despite withholding enteral feedings for VLBW babies for
years, there are also studies supporting EEF of VLBW babies,
showing that these babies have a decreased incidence of
sepsis and shorter length of stay (Flidel-Rimon et al., 2004).
Early enteral feeding policy has been adopted in our NICU
since October 2004. This study was conducted to answer the
question whether EEF increases the risk of NEC in VLBW

Does EEF increase the risk of NEC

FC Celik et al

581
babies and whether early and late feeding policies have any
impact on daily weight gain, length of stay and mortality.

Patients and methods

The unit where the study was conducted was a Level III
NICU. Following the local ethics committees approval (no.
298), the data of all VLBW babies followed in the NICU
between April 2003 and April 2006 was retrospectively
analyzed from the NICU database and patient files. During
this period, of the 300 VLBW babies followed in the unit,
three were term small for gestational age babies, and hence
were excluded and the data were analyzed on 297 preterm
VLBW babies.
The study period was divided into two phases: the first
phase was between April 2003 and October 2004, when
babies were not fed enterally until they were extubated. One
hundred and sixty-seven VLBW preterm babies were followed in that period. In November 2004, following the 21st
International Symposium on Neonatal Intensive Care and
lectures on early aggressive nutrition of preterm babies, the
feeding policy of the unit was changed to EEF, which could
be summarized as feeding VLBW babies as soon as possible
(preferably starting from the first day of life) and even when
they were intubated. So, the second phase was between
November 2004 and April 2006. One hundred and thirty
preterm VLBWs were followed during that period. Criteria
for withholding enteral feedings in both periods were the
same (except endotracheal intubation): hypotension necessitating vasopressor agent use (dopamine or dobutamin45 mg per kg per min), abdominal distention,
abdominal tenderness and suspected or proven NEC.
Enteral feeding practices were also the same in both
periods: enteral feedings were started preferably with breast
milk (BM). When BM was not available, standard preterm
formula was used. One milliliter of BM/preterm formula was
given four times on the first day of enteral feeding, and 8
times on the second day. This was increased to not more
than 24 ml per kg per day on the consecutive days. Babies
getting solely BM were given human milk fortifier (Eoprotin)
when they were able to get 450 ml kg1 BM enterally.
Intubated and mechanically ventilated babies were fed via
an orogastric tube, by boluses.
Total parenteral nutrition with 1 g per kg per day protein
and 1 g per kg per day lipid was started on the morning of
admission to the NICU during the whole study period.
Protein content of total parenteral nutrition was increased to
3 g per kg per day and lipid to 2 g per kg per day with 1 g kg1
daily increments. All total parenteral nutrition solutions
were prepared in a separate total parenteral nutrition unit.
Parenteral nutrition was discontinued when 100 ml per kg
per day enteral feeding was achieved. For every baby, the
time for first enteral feeding, presence of NEC, daily weight
gain and length of hospital stay were recorded in addition to
gestational age, birth weight and way of birth. The possible

risk factors for NEC were also recorded, which were umbilical
vein catheterization, small for gestational age birth, polycythemia, partial exchange transfusion for polycythemia,
sepsis (clinical or blood culture proven), patent ductus
arteriozus and antenatal corticosteroid use (as a protective
factor).
Necrotizing enterocolitis was classified according to modified Bells staging criteria (Walsch and Klegman, 1986). Stage
Ia and Ib cases of NEC were treated with parenteral
antibiotics and bowel rest for 37 days. Stage II or greater
cases were treated with parenteral antibiotics and bowel rest
for 47 days.
Patent ductus arteriozus was diagnosed by echocardiogram
and treated with oral ibuprophen.
Partial exchange transfusion was performed in polycythemic babies whenever hypoglycemia and/or thrombocytopenia and/or poor signs of peripheral perfusion occurred and
when there was no other reason than polycythemia to
explain these signs.
llners
Clinical sepsis was diagnosed according to To
scoring system. Babies getting 10 points and higher were
diagnosed as clinical sepsis (Tollner, 1982). Every baby
diagnosed as clinical sepsis underwent sepsis work-up and
was treated with appropriate antibiotics.
According to the NICU protocol, none of the babies had
H2-receptor blocker therapy or corticosteroid use before 28
days of life during the whole study period.

Statistical analysis
The Statistical Package for the Social Sciences (SPSS) for
Windows (SPSS Inc., Chicago, IL, USA), version 13.0, was
used for statistical analysis. Data were recorded as mean
values (s.d. and minimummaximum) and percentages. The
significance of mean differences between groups was
assessed by MannWhitney U-test.
Categorical variables were compared using the w2-test and
Fishers exact test.
Multivariate logistic regression analysis was used to
identify the risk factors associated with NEC. When NEC
was the dependent variable, the independent variables were
gestational age, birth weight, sex, polycythemia, partial
exchange transfusion, umbilical vein catheterization, small
for gestational age birth, positive blood culture, sepsis,
patent ductus arteriozus, antenatal corticosteroid use and
endotracheal intubation. Independent variables that had a Pvalue of o0.05 were retained into a regression model. The
results of logistic regression analyses were expressed as odds
ratios (OR) with 95% confidence intervals (CI). A P-value of
o0.05 was considered statistically significant.

Results
Clinical characteristics of the babies and risk factors for NEC
in early and late feeding periods are shown in Table 1. The
European Journal of Clinical Nutrition

Does EEF increase the risk of NEC

FC Celik et al

582
Table 1 Clinical characteristics of the babies and risk factors for NEC in early and late feeding periods
Late feeding (n 167)
Birth weight (g)
Gestational age (weeks)
Female/male (%)
Antenatal corticosteroid use (%)
SGA birth (%)
Polycythemia (%)
Partial exchange transfusion (%)
Patent ductus arteriozus (%)
Respiratory distress syndrome (%)
Endotracheal intubation (%)
Umbilical vein catheterization (%)
Clinical sepsis (%)
Positive blood culture (%)
First enteral feeding (days)
Necrotizing enterocolitis (%)
Average weight gain (g day1)
Length of stay (day)
Death (%)

1130.5243.6
28.52.5
97/70
41
25
13
2
14
64
111
20
48
18
9.010.0
12
12.88.4
27.323.1
52

Early feeding (n 130)

(4901495)
(2236)
(58.1/41.9)
(24.6)
(15.0)
(7.8)
(1.2)
(8.4)
(38.3)
(66.5)
(11.9)
(28.7)
(10.8)
(147)
(7.2)
(14, 40)
(1109)
(31.1)

1151.3236.5
29.32.7
66/64
31
26
16
6
17
44
82
105
52
27
4.44.6
8
16.310.9
31.825.5
34

(6001495)
(2236)
(50.8/49.2)
(23.8)
(20.0)
(12.3)
(4.6)
(13.1)
(33.8)
(63.1)
(80.8)
(40)
(20.8)
(123)
(6.2)
(13, 42)
(1147)
(26.1)

P-value
40.05
40.05
40.05
40.05
40.05
40.05
40.05
40.05
40.05
40.05
o0.001*
0.003*
0.03*
o0.001*
40.05
0.011*
40.05
40.05

Abbreviations: NEC, necrotizing enterocolitis; SGA, small for gestational age.

Values are given as averages.d. (minimummaximum).
*Po0.05.

average birth weight, gestational age, antenatal corticosteroid use, small for gestational age birth and babies diagnosed
with respiratory distress syndrome were similar in both
groups. Clinical characteristics of the patients with NEC,
according to early and late feeding regimens, are shown in
Table 2. Twenty (6.7%) patients during the study period were
diagnosed as NEC. None had gastrointestinal abnormalities.
Risk factors associated with NEC are shown in Table 3.
Following logistic regression analysis, the main risk factors
found to be associated with NEC were sepsis (OR: 5.5, 95%
CI: 1.225.7) and blood culture positivity (OR: 3, 95% CI:
1.18.1).

Discussion
In the present study, there was no difference in the incidence
of NEC between early and late enteral fed groups, both
including babies vulnerable to NEC, with similar average
gestational age and birth weight. The present finding that
EEF does not increase the risk of NEC was consistent with the
results of McClure (2001), Davey et al. (1994) and FlidelRimon et al. (2004), who recommend EEF of VLBW babies
and report positive outcomes with EEF. Enteral feedings
improve the activity of digestive enzymes, enhance the
activity of digestive hormones and increase the intestinal
blood flow and motility in preterms. Infants given early
trophic feeds were found to have better feeding tolerance,
improved growth, reduced length of hospitalization and
decreased likelihood of sepsis as compared with infants who
are not (McClure, 2001; Ziegler et al., 2002).
In the study of Flidel-Rimon et al., EEF starting at the
second or third day of life appeared to be associated with a
European Journal of Clinical Nutrition

reduced risk of nosocomial sepsis without incurring an

increased risk of NEC (Flidel-Rimon et al., 2004). This finding
was also confirmed in the studies of McClure (2001) and
Newell (2000); that is, theearly trophic feeds do not increase
susceptibility to NEC.
In contrast to the study by Flidel-Rimon et al., in our study
early enteral fed babies had a higher incidence of clinical
sepsis and blood culture positivity than late fed ones. Despite
an increase in clinical sepsis and blood culture positivity, the
incidence of NEC, length of stay in NICU and mortality
ratios did not differ significantly between the two feeding
regimens.
Besides, although sepsis and blood culture positivity were
the most important risk factors for NEC, the ratio of babies
with these two entities were similar in babies with NEC, both
in the early and late fed groups. We think that the data on
the increased risk for sepsis would be interpreted cautiously
before arguing that EEF was responsible for that. Umbilical
vein catheterization, which is a well-known risk factor for
bloodstream infections in newborns (Moro et al., 1996), had
increased sevenfold, which might have influenced the
increased risk for sepsis, but without an impact on NEC risk.
On logistic regression analysis, the most important risk
factors found to be associated with NEC were sepsis
(Po0.001) and blood culture positivity (P 0.001). This
result was consistent with the literature (Krediet et al., 2003).
We could not show the well-known protective effect of
antenatal corticosteroids on NEC, probably due to the low
ratio of antenatal steroid use in our patients.
In the study of Khayata et al. (1987), the mean weight gain
was 70 g week1 in the early fed and 71 g week1 in the late
fed VLBW babies, without a significant difference. In our
study, the average daily weight gain of patients increased

Does EEF increase the risk of NEC

FC Celik et al

583
Table 2 Clinical characteristics of patients with NEC
Late feeding (n 12)
1059.1251.9
28.73.3
9/3
2
3
3
4
10.610.4
7
1
4
5
5
12.24.3
40.026.0
38.3 26.3
42.4 28.4
7

Birth weight (g)

Gestational age (weeks)
Female/male (%)
490749 g (%)
750999 g (%)
10001249 g (%)
12501499 g (%)
First enteral feeding (days)
Stage I NEC (%)
Stage II NEC (%)
Stage III NEC (%)
Clinical sepsis (%)
Positive blood culture (%)
Average weight gain (g day1)
Length of stay (days)
Length of stay in babies who died
Length of stay in babies surviving NEC
NEC-related death (%)

(7201450)
(2334)
(75/25)
(16.7)
(25)
(25)
(33.3)
(337)
(58.3)
(8.3)
(33.3)
(41.7)
(41.7)
(618)
(883)
(880)
(1483)
(58.3)

Early feeding (n 8)
1095.0225.6
28.52.9
4/4
0
4
2
2
6.08.4
3
1
4
4
4
13.98.4
51.620.1
51.5 43.1
51.7 13.9
2

(8701470)
(2534)
(50/50)
(0)
(50)
(25)
(25)
(121)
(37.5)
(12.5)
(50)
(50)
(50.0)
(530)
(2182)
(2182)
(2565)
(25)

P-value
40.05
40.05
40.05

40.05

40.05
40.05
40.05
40.05
40.05
40.05
40.05

Abbreviation: NEC, necrotizing enterocolitis.

Values are given as averages.d. (minimummaximum).

Table 3 Risk factors for NEC

NEC ( ) (n 20)
Gestational age (weeks)
Birth weight (g)
Male/Female (%)
Polycythemia (%)
Partial exchange transfusion (%)
Umbilical vein catheters (%)
SGA birth (%)
Positive blood culture (%)
Clinical sepsis (%)
Patent ductus arteriozus (%)
Antenatal corticosteroid use (%)
Endotracheal intubation (%)

28.63.1
1073.5236.2
7/13
2
1
8
6
9
18
5
3
13

(2334)
(7201470)
(35/65)
(10)
(5)
(40)
(30)
(45)
(90)
(25)
(15)
(65)

NEC () (n 277)

28.82.5
1145.9243.0
129/148
23
5
98
45
30
114
26
61
150

(2236)
(4901495)
(46.6/53.4)
(8.3)
(1.8)
(35.4)
(16.2)
(10.8)
(41.1)
(9.4)
(22)
(54.1)

P-value
40.05
40.05
40.05
40.05
40.05
40.05
40.05
0.001*
o0.001*
40.05
40.05
40.05

Abbreviations: NEC, necrotizing enterocolitis; SGA, small for gestational age.

Values are given as averages.d. (minimummaximum).
*Po0.05.

significantly after the early feeding protocol (89.6 versus

114.1 g week1). Also, the weight gain in our study far
exceeded the weight gain of the babies reported by Khayata
et al. (1987).
The goal of postnatal feeding strategies is to reach the
intrauterine growth rate of 1015 g per kg per day (Puntis,
2006), which is not always possible to achieve in sick VLBW
babies.
Other studies have also shown that EEF significantly
increases average daily weight gain, which is related to
better catch-up growth and better outcomes on neurodevelopment and linear growth (Ehrenkranz et al., 2006). To put
it another way, by feeding the preterm baby, we are also
feeding his brain. This should be an important reason to feed
babies more aggressively and earlier. Tyson et al. (2007) also
emphasize that the ideal feeding regimen of a VLBW baby

should be the one that increases survival without neurodevelopmental impairment.

Mortality rates attributed to NEC have been reported at
1530% (Pietz et al., 2007). In the present study, the ratio of
babies dying of NEC was 58.3% in the late fed group, whereas
it was 25% in the EEF group. Since NEC was treated similarly
during the two study periods in the unit, although not
statistically significant, and the number is small to conclude,
the decrease in mortality might be due to the positive effects
of EEF on the immature intestine. Enterocytes rely on the
luminal contents for growth and nutrition. EEF might have
prevented severe atrophy in intestine developing within 3
days of starvation, as shown in animal studies (Hughes and
Doweling, 1980).
Cochrane cautiously approaches early trophic feedings of
VLBW infants by stating that even when trophic feedings
European Journal of Clinical Nutrition

Does EEF increase the risk of NEC

FC Celik et al

584
were compared to no feedings, the relative risk for NEC was
1.16 (0.751.79) (Tyson and Kennedy, 2005). Contrarily,
the incidence of NEC did not increase in the EEF group
in our study. This result is also supported by the study by
Sisk et al. (2007), who fed 72% of 202 VLBW babies in
the first three days of life and 97% in the first week. An
enteral feeding of 50 ml per kg per day has been achieved
on an average of 11.2 days and 100 ml per kg per day on
16.3 days. Their overall NEC rate was 5%, similar to ours,
and still lower than the 7% NEC ratio of the NICHD
Neonatal Network data. We hope that other NICUs adopting
EEF protocols and publishing their results will provide
better information on EEF of VLBW babies in a positive or
negative way.
According to our results, the EEF of VLBW babies does not
increase the incidence of NEC. The superiority of BM over
formula on the incidence of infection in VLBW babies has
not been proved in systematic reviews (De Silva et al., 2004).
Contrarily, in a recent study, Sisk et al. (2007) have shown
that enteral feeding containing at least 50% human milk in
the first 14 days of life was associated with a sixfold decrease
in the risk of NEC.
In preterm babies in an intensive care unit, an abnormal
pattern of bowel colonization occurs when compared with
that of healthy term newborns. Bifidobacteria appears on the
11th day and is predominant in 20 days in breast-fed VLBW
infants. This is 4 days in term bread fed babies (Sakata et al.,
1985). Enteral feeding in the first 2 weeks with BM reduces
inflammation by promoting symbiotic bacterial colonization
and/or through other anti-inflammatory properties that it
possesses, and it may decrease the incidence of NEC (De Silva
et al., 2004). We conclude that EEF of VLBW babies should
not be withheld for fear of NEC, and BM would preferably be
the first feed to reach the vulnerable intestine.

Acknowledgements
kru
Ku
cu
ko
du
k, MD, for his support of
We thank Professor Su
all our scientific work and Eilon Shany, MD, for his
encouragement on early enteral feeding.

References
Davey AM, Wagner CL, Cox C, Kendig JW (1994). Feeding premature
infants while low umbilical artery catheters are in place: a
prospective, randomized trial. J Pediatr 124, 795799.

European Journal of Clinical Nutrition

De Silva A, Jones PW, Spencer SA (2004). Does human milk reduce

infection rates in preterm infants? A systematic review. Arch Dis
Child Fetal Neonatal Ed 89, F509F513.
Ehrenkranz RA, Dusick AM, Vohr BR, Wright LL, Wrage LA, Poole WK
(2006). Growth in the neonatal intensive care unit influences
neurodevelopmental and growth outcomes of extremely low birth
weight infants. Pediatrics 117, 12531261.
Flidel-Rimon O, Friedman S, Lev E, Juster-Reicher A, Amitay M,
Shinwell ES (2004). Early enteral feeding and nosocomial sepsis in
very low birth weight infants. Arch Dis Child FNE 89, F289F292.
Guillet R, Stoll BJ, Cotten CM, Gantz M, McDonald S, Poole WK et al.
(2006). National Institute of Child Health and Human Development Neonatal Research Network. Association of H2-blocker
therapy and higher incidence of necrotizing enterocolitis in very
low birth weight infants. Pediatrics 117, 137142.
Hughes CA, Doweling RH (1980). Speed of onset of adaptive mucosal
hypoplasia and hypofunction in the intestine of parenterally fed
rats. Clin Sci 59, 317327.
Khayata S, Gutcher G, Bamberger J (1987). Early versus late feeding of
low birth weight (LBW) infants: effect on growth and hyperbilirubinemia. Pediatr Res 21, 431A.
Krediet TG, van Lelyveld N, Vijlbrief DC, Brouwers HA, Kramer WL,
Fleer A et al. (2003). Microbiological factors associated with
neonatal necrotizing enterocolitis: protective effect of early
antibiotic treatment. Acta Paediatr 92, 11801182.
Lin PW, Stoll BJ (2006). Necrotizing enterocolitis. Lancet 368,
12711283.
McClure RJ (2001). Trophic feeding of the preterm infant. Acta
Paediatr Suppl 90, 1921.
McKeown RE, Marsh TD, Amarnath U, Garrison CZ, Addy CL,
Thompson SJ et al. (1992). Role of delayed feeding and of feeding
increments in necrotizing enterocolitis. J Pediatr 121, 764770.
Moro ML, De Toni A, Stolfi I, Carrieri MP, Braga M, Zunin C (1996).
Risk factors for nosocomial sepsis in newborn intensive and
intermediate care units. Eur J Pediatr 155, 315322.
Newell SJ (2000). Enteral feeding of the micropremie. Clin Perinatol
27, 221234.
Pietz J, Achanti B, Lilien L, Stepka EC, Mehta SK (2007). Prevention of
necrotizing enterocolitis in preterm infants: a 20-year experience.
Pediatrics 119, e164e170.
Puntis JW (2006). Nutritional support in the premature newborn.
Postgrad Med J 82, 192198.
Sakata H, Yoshika H, Fujita K (1985). Development of the intestinal
flora in very low birth weight infants compared to normal fullterm newborns. Eur J Pediatr 114, 186190.
Sisk PM, Lovelady CA, Dillard RG, Gruber KJ, OShea TM (2007).
Early human milk feeding is associated with a lower risk of
necrotizing enterocolitis in very low birth weight infants. J
Perinatol 27, 428433.
Tollner U (1982). Early diagnosis of septicemia in the newborn.
Clinical studies and sepsis score. Eur J Pediatr 138, 331337.
Tyson JE, Kennedy KA (2005). Trophic feedings for parenterally fed
infants. Cochrane Database Syst Rev 20, CD000504.
Tyson JE, Kennedy KA, Lucke JF, Pedroza C (2007). Dilemmas
initiating enteral feedings in high risk infants: how can they be
resolved? Semin Perinatol 31, 6173.
Walsch MC, Klegman RM (1986). Necrotizing enterocolitis: treatment based on staging criteria. Pediatr Clin North Am 33, 179201.
Ziegler EE, Thureen PJ, Carlson SJ (2002). Aggressive nutrition of the
very low birth weight infant. Clin Perinatol 29, 224225.

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.