2. 2 Journal of Parenteral and Enteral Nutrition XX(X)
Table 1. Nutrition Support Guideline Recommendations for Neonatal Patients at Risk for Necrotizing Enterocolitis
Question Recommendation Grade
When and how should feeds be started in We suggest that minimal enteral nutrition should be initiated Weak
infants at high risk for NEC? within the first 2 days of life and advanced by 30 mL/kg/d in
infants ≥1000 g.
Does the provision of mother’s milk reduce We suggest the exclusive use of mother’s milk rather than bovine- Weak
the risk of developing NEC relative to based products or formula in infants at risk for NEC.
bovine-based products or formula?
Do probiotics reduce the risk of developing There are insufficient data to recommend the use of probiotics in Further research
NEC? infants at risk for NEC. needed
Do certain nutrients either prevent or We do not recommend glutamine supplementation for infants at Strong
predispose to the development of NEC? risk for NEC.
There is insufficient evidence at this time to recommend arginine Further research
and/or long-chain polyunsaturated fatty acid supplementation needed
for infants at risk for NEC.
When should feeds be reintroduced to infants There are insufficient data to make a recommendation regarding Further research
with NEC? time to reintroduce feedings to infants after NEC. needed
Abbreviation: NEC, necrotizing enterocolitis.
support in all healthcare settings. These Clinical Guidelines for the body of evidence and for the recommendation. The pro-
were developed under the guidance of the A.S.P.E.N. Board of cedures listed below were adopted from the GRADE process for
Directors. Promotion of safe and effective patient care by use with A.S.P.E.N. Clinical Guidelines with consideration of
nutrition support practitioners is a critical role of the A.S.P.E.N. the levels of review (by internal and external content reviewers,
organization. The A.S.P.E.N. Board of Directors has been pub- by the A.S.P.E.N. Board of Directors).
lishing Clinical Guidelines since 1986.12-23 The A.S.P.E.N. A rigorous literature search is undertaken to locate clinical
Clinical Guidelines editorial board evaluates in an ongoing outcomes associated with practice decisions in the population
process when individual Clinical Guidelines should be of interest. Each pertinent paper is appraised for evidence qual-
updated. ity according to research quality (randomization, blinding,
These A.S.P.E.N. Clinical Guidelines are based upon gen- attrition, sample size, and risk of bias for clinical trials24 and
eral conclusions of health professionals who, in developing prospective vs retrospective observation, sample size, and
such guidelines, have balanced potential benefits to be derived potential bias for observational studies) and placed into an evi-
from a particular mode of medical therapy against certain risks dence table. A second table is used to provide an overview of
inherent with such therapy. However, the professional judg- the strength of the available evidence according to the clinical
ment of the attending health professional is the primary com- outcomes, in order to support a consensus decision regarding
ponent of quality medical care. Because guidelines cannot the guideline recommendation. If the evidence quality is high,
account for every variation in circumstances, the practitioner it is unlikely that further research will change our confidence
must always exercise professional judgment in their applica- in the estimate of effect. With moderate grade evidence, further
tion. These Clinical Guidelines are intended to supplement, but research is likely to modify the confidence in the effect esti-
not replace, professional training and judgment. mate and may change the estimate. With low grade evidence,
A.S.P.E.N. has adopted concepts of the Grading of further research is very likely to change the estimate, and with
Recommendations, Assessment, Development and Evaluation very low evidence quality, the estimate of the effect is very
(GRADE) working group (http://www.gradeworkinggroup.org) uncertain. A clinical recommendation is then developed by
for development of its clinical guidelines. The GRADE working consensus of the Clinical Guidelines authors, based on the best
group combined the efforts of evidence analysis methodologists available evidence. The risks and benefits to the patient are
and clinical guidelines developers from diverse backgrounds weighed in light of the available evidence. Conditional lan-
and health organizations to develop an evaluation system that guage is used for weak recommendations. For further details
would provide a transparent process for evaluating the best on the A.S.P.E.N. application of GRADE, see the “Clinical
available evidence and integration of the evidence with clinical Guidelines for the Use of Parenteral and Enteral Nutrition in
knowledge and consideration of patient priorities. These proce- Adult and Pediatric Patients: Applying the GRADE System to
dures provide added transparency by developing separate grades Development of A.S.P.E.N. Clinical Guidelines.”24
Downloaded from pen.sagepub.com by guest on August 24, 2012
3. A.S.P.E.N. Clinical Guidelines / Fallon et al 3
For the current Clinical Guideline, the search term necrotiz- instability can impact feeding practices, and thus, discretion
ing enterocolitis was used in PubMed with inclusion criteria should be employed under these circumstances. Last, one
including infants (birth to 23 months); humans; clinical trial; RCT26 evaluated the effect of stable (20 mL/kg/d) vs advancing
randomized controlled trial; case reports; clinical trial: phase I, (20 mL/kg/d to goal 140 mL/kg/d) feeding volumes for a
phase II, phase III, phase IV; comparative study; controlled 10-day period following the initiation of enteral nutrition and
clinical trial; guideline; journal article; multicenter study; found a significantly higher incidence of NEC in infants fed
English language; and published within the last 10 years. The advancing volumes. Due to the high incidence of NEC in the
search was conducted on April 21, 2011. The questions are advancing group (10% vs 1.4%), the study was prematurely ter-
summarized in Table 1. For questions 1 and 3, an additional minated. It is important to recognize that a major difference in
limitation of randomized controlled trial was implemented due this study compared with the previous RCTs29-31 is that enteral
to the plethora of literature on these topics. For questions 2, 4, nutrition was initiated later in life, the timing of which was at
and 5, pertinent literature within the past 10 years, without the discretion of the neonatologist. Specifically, the earliest age
restriction to evidence type, was included. A total of 1335 of feed initiation was 4 days with a median age of 9.5 days in
abstracts were reviewed, of which 24 papers met the inclusion infants who developed NEC and 11 days for infants in the
criteria of the Clinical Guidelines and were included. advancing group who developed NEC. This is a potentially
important confounding variable, making the interpretation of
these study results complicated and to be taken with caution.
Practice Guidelines and Rationales Although the majority of these aforementioned studies have
Question 1: When and how should feeds be started in recommended larger, multicentered prospective trials to fur-
infants at high risk for NEC (Tables 2 and 3)? ther evaluate questions on enteral nutrition initiation and
Recommendation: We suggest that minimal enteral advancement, based on the available data, early MEF within
nutrition be initiated within the first 2 days of life and the first 2 days of life and advancement at 30 mL/kg/d in
advanced by 30 mL/kg/d in infants ≥1000 g. infants ≥1000 g can be suggested.
Grade: Weak
Rationale: Several randomized controlled trials (RCTs) Question 2: Does the provision of mother’s milk reduce
have been conducted to gain insight into the optimal time of the risk of developing NEC relative to bovine-based prod-
initiation and rate of advancement of enteral nutrition in ucts or formula (Tables 4 and 5)?
infants at risk for NEC. Of the studies reviewed, 2 of 1025,26 Recommendation: We suggest the exclusive use of
evaluated NEC (Bell’s stage ≥II) as the primary outcome, mother’s milk rather than bovine-based products or formula
whereas the remaining studies predominantly evaluated feed- in infants at risk for NEC
ing tolerance and/or time to achievement of full enteral nutri- Grade: Weak
tion with NEC as a secondary outcome measure. With regard Rationale: The type of enteral nutrition administered to
to the timing of initiation of EN, one RCT25 evaluated the an infant at risk for NEC is important. Several studies have
effect of early (≤5 days; median 2 days) vs delayed (≥6 days; focused on whether the administration of human milk results in
median 7 days) initiation of minimal enteral feeding (MEF) in a reduction in the incidence of NEC compared with formula
infants with an age-adjusted birth weight (BW) ≤10th percen- feeding. These studies used mother’s milk (MM) with the
tile and intrauterine growth restriction (IUGR). No difference exception of one,32 which fed infants pasteurized donor milk
in the incidence of NEC between groups was found, although (DM) if MM was unavailable. Exclusive feeding with MM is
it was concluded that a larger sample size would be needed to associated with a decreased risk of NEC as compared with pre-
adequately evaluate for an effect. Two RCTs27,28 evaluated the term formula (PF).33 As a substitute for MM, pasteurized DM
effect of MEF vs nil per os (NPO) status within the first week has not been found to have any protective effect over PF with
of life with feeds beginning at a median age of 2 days in regard to the incidence of NEC,34 but feeding with MM and/or
<1000-g and <2000-g infants, respectively, and found no sig- DM has been found to be associated with a decreased risk of
nificant differences in the incidence of NEC. For these studies, NEC as compared with a combination of MM and/or DM and
the quantity associated with MEF was ≤12 mL/kg/d. bovine milk (BOV)–based products.32 It is important to note
Three RCTs29-31 evaluated the effect of slow (15–20 mL/ that the source of enteral nutrition was explicitly evaluated;
kg/d) vs rapid (30 mL/kg/d) enteral nutrition advancement to a supplementation with milk fortifier was not evaluated. With
goal rate of between 150 and 180 mL/kg/d and found that rapid respect to the quantity of MM administered, one prospective
advancement was well tolerated by infants with an average BW cohort study35 found enteral nutrition with ≥50% MM within
1000–2000 g without an increased incidence of NEC. In these the first 14 days was associated with a 6-fold decreased risk of
studies, enteral nutrition was initiated at a median age of 6 NEC. An observational study36 found the amount of daily MM
hours29 or 2 days.30,31 However, these studies were not powered (1 to ≥50 mL/kg/d) fed through week 4 of life had no effect on
to detect statistically significant differences in the incidence of the incidence of NEC. Based on the available data, exclusively
NEC. In addition, it is important to note that hemodynamic fed MM has been shown to be beneficial in the prevention of
Downloaded from pen.sagepub.com by guest on August 24, 2012
4. 4 Journal of Parenteral and Enteral Nutrition XX(X)
NEC and is therefore recommended over formula feeding. It is others may actually predispose infants to NEC. With
unclear whether the amount and/or timing of MM administered respect to amino acids (AA), recent literature has focused on
have an effect on the incidence of NEC, and therefore no rec- the effect of arginine and glutamine supplementation on the
ommendation regarding the optimal dose of MM can be made. incidence of NEC. The plasma arginine and asymmetric
dimethylarginine (ADMA, a metabolic by-product of pro-
Question 3: Do probiotics reduce the risk of developing tein modification processes) concentrations as well as the
NEC (Tables 6 and 7)? arginine:ADMA ratio have been found to be lower in prema-
Recommendation: There are insufficient data to recom- ture infants with NEC, which have subsequently been shown
mend the use of probiotics in infants at risk for NEC. to be associated with an increased mortality.44 Although
Grade: Further research needed there is not an abundance of literature, one RCT45 focuses on
Rationale: There has been much debate over the admin- the effect of prophylactic L-arginine supplementation (1.5
istration of oral probiotics in the prevention of NEC. Seven mmol/kg/d), the results of which suggest that supplementa-
RCTs evaluating the use of prophylactic probiotics in preterm tion may be effective in reducing the overall incidence of
and very low birth weight (VLBW) infants met inclusion crite- NEC. Of note, the results of this study must be taken with
ria for these guidelines. NEC, defined as Bell’s stage ≥II, was caution as the sample size was small and results demon-
the primary end point in 6 of 7 studies.37-42 The type of bacte- strated no difference in the reduction of Bell’s stage ≥II.
ria, dosage, frequency, and duration of treatment varied widely Glutamine supplementation has additionally been evaluated
across studies. One study evaluated the effect of a probiotic in one large, well-conducted RCT, and no statistically sig-
with MM vs MM alone,40 whereas 2 studies evaluated the nificant difference in the incidence of NEC between supple-
effect of a probiotic with human milk (HM; MM or DM) vs mented and nonsupplemented groups was found.46 Apart
HM alone,37,42 and 3 studies evaluated the effect of a probiotic from individual amino acid supplementation, the administra-
with MM or formula vs MM or formula alone.39,41,43 One tion of AA in parenteral nutrition (PN) has additionally been
study38 specifically compared the effect of killed probiotic studied. In a comparative pre- and postintervention study,47
(KP) vs living probiotic (LP) Lactobacillus acidophilus on the early AA administration was associated with an increased
incidence of NEC and found no difference in the incidence of incidence of surgical NEC in VLBW infants. Last, fatty acid
NEC between groups; LP and KP were both found to be pre- supplementation was evaluated in one RCT48; this study
ventative against NEC in comparison to a placebo group, with demonstrated a slightly increased incidence of NEC (5.3%
KP retaining similar benefits to live bacteria with no adverse vs 2%) in long-chain polyunsaturated fatty acid (LCPUFA)–
effect. All 7 RCTs demonstrated a lower incidence of NEC in supplemented (fat mixture containing linoleic, α-linolenic,
the group of infants who received probiotics as compared with and γ-linolenic acids) compared with nonsupplemented
those who did not receive probiotics, although 1 of 7 studies43 infants, although the difference between groups was not sta-
did not demonstrate statistical significance between groups. tistically significant. Based on the aforementioned studies,
Although the implementation of a “probiotic” resulted in a there is limited research on AA/fatty acid administration and
lower incidence of NEC across studies, it is important to supplementation, and it remains an important area for future
emphasize that these studies used different types of probiotics, research. However, based on the available literature, argi-
with some administering a combination of probiotics.37,39-42 It nine supplementation may be effective, although the evi-
is additionally important to mention that there is no Food and dence is underpowered, whereas glutamine supplementation
Drug Administration (FDA) approval to date for routine use of does not appear to prevent NEC, and LCPUFA supplementa-
these products. Further studies are necessary to determine the tion may predispose infants to NEC. Therefore, although
most effective type(s) of probiotic, dosage, and duration of glutamine supplementation is not recommended for infants
treatment; thus, no recommendation on the use of probiotics in at risk for NEC, there is insufficient evidence to recommend
infants at risk for NEC can be made at this time. arginine and/or LCPUFA supplementation at this time.
Question 4: Do certain nutrients either prevent or predis- Question 5: When should feeds be reintroduced to
pose to the development of NEC (Tables 8 and 9)? infants with NEC (Tables 10 and 11)?
Recommendation: We do not recommend glutamine Recommendation: There are insufficient data to make a
supplementation for infants at risk for NEC. There is insuf- recommendation regarding time to reintroduce feedings to
ficient evidence at this time to recommend arginine and/or infants after NEC.
long-chain polyunsaturated fatty acid supplementation for Grade: Further research needed
infants at risk for NEC. Rationale: There is no standard recommendation as to
Grade: Strong (glutamine); further research needed when enteral nutrition should be reinitiated after a defini-
(arginine, long-chain polyunsaturated fatty acids) tive diagnosis of NEC. Historically, it has been suggested that
Rationale: A review of the literature suggests that cer- a period of fasting from 10 days to 3 weeks should be observed
tain nutrients may reduce the incidence of NEC, whereas prior to the reintroduction of enteral nutrition for an infant with
Downloaded from pen.sagepub.com by guest on August 24, 2012
5. A.S.P.E.N. Clinical Guidelines / Fallon et al 5
NEC; however, these recommendations are not founded on catheter-related sepsis and post-NEC intestinal stricture for-
scientific data. It is without question that practices vary greatly mation as well as shorter time to full enteral nutrition and
among institutions and physicians. It has been suggested that shorter hospitalization. It is important to note that both studies
prolonged fasting may actually be detrimental due to the have serious limitations given their retrospective nature, and
potential need for prolonged central venous access and PN, both are underpowered to assess the impact of early enteral
and has prompted some institutions to introduce early feeding nutrition on NEC recurrence, a very important outcome vari-
regimens in infants with NEC. To date, the literature on the able. Although these studies suggest that prolonged fasting
impact of early feeding regimens is limited. In fact, only 2 ret- periods traditionally recommended for infants with NEC may
rospective reviews met the inclusion criteria for these guide- not be necessary, further prospective and randomized con-
lines. Both of these studies suggest that, for infants with trolled trials are necessary before recommendations can
Bell’s stage II NEC, early feeding regimens may actually have be made as to when feeds should be reintroduced in infants
potential beneficial effects, including a reduced incidence of with NEC.
(Text continues on p. 17.)
Downloaded from pen.sagepub.com by guest on August 24, 2012
6. Table 2. Evidence Table Question 1: When and how should feeds be started in infants at high risk for NEC?
6
Study Design,
Author, Year Quality Population, Setting, N Study Objective Results Comments
Krishnamurthy,29 RCT Preterm infants <34 weeks To evaluate the effect of slow (20 Incidence of NEC: 2% Rapid enteral nutrition advancement
2010 Nonblinded to GA with BW 1000–1499 mL/kg/d) vs rapid (30 mL/kg/d) (1/50) in the slow- of 30 mL/kg/d is well tolerated
intervention g, born 2/2008 to 9/2008, enteral nutrition advancement feeding advancement without an increased incidence of
(investigators) and followed until they had by nasogastric bolus on the time group and 4% (2/50) NEC in stable preterm neonates
Rate of attrition: regained BW or developed to achievement of full enteral in the rapid-feeding weighing 1000–1499 g.
10/100 NEC nutrition (180 mL/kg/d) advancement group (P There was no statistically significant
Tertiary care hospital (India) Primary outcome: Time to attainment = 1.0) difference in mortality between
N = 100 (n = 50/group) of full enteral nutrition groups.
Secondary outcome: Incidence The study was not powered to detect
of feeding intolerance, NEC clinical or statistical differences in
(Bell’s stage ≥IIA), mortality, the incidence of NEC as NEC was
apnea, duration of hospital stay/ not the primary outcome.
intravenous fluids, weight gain,
and nosocomial sepsis
Karagianni,25 RCT Preterm infants 27–34 weeks To examine the effect of early Incidence of NEC: 15% Early MEF for preterm infants
2010 Nonblinded GA with age-adjusted BW (≤5 days) vs delayed (≥6 days) (6/40) in the early MEF with IUGR and abnormal
Pilot study ≤10th percentile (IUGR), initiation of MEF on the incidence group and 9.8% (4/41) antenatal Doppler results does not
Rate of attrition: Apgar score >5, and arterial of NEC and feeding intolerance in the delayed MEF significantly impact the incidence
3/84 cord blood pH ≥7.0 with Primary outcomes: Incidence of group (RR = 1.54 early of NEC.
abnormal antenatal Doppler NEC (Bell’s stage ≥II) and feeding MEF; 95% CI 0.469– Mortality was not significantly
results, admitted between intolerance 5.043) different between groups (P =
5/2004 and 5/2008 Secondary outcome: Mortality .512).
Level III NICU (Greece)
Downloaded from pen.sagepub.com by guest on August 24, 2012
N = 84 (n = 42/group)
Mosqueda,27 RCT ELBW infants (BW ≤1000 g), To evaluate the efficacy of early Incidence of NEC: 9% There was no difference in the
2008 Nonblinded admitted between 1/2001 MEF (12 mL/kg/d) on overall (3/33) in the MEF group incidence of NEC between MEF
Pilot study and 8/2003. Infants were feeding tolerance in infants from and 14% (4/28) in the and NPO groups.
Rate of attrition: separated into MEF and DOL 2–7 NPO group (P = .53) Based on the incidence of NEC
23/84 NPO groups from DOL Primary outcome: Feeding tolerance Mortality was 17% in the and mortality in this study, a
2–7. On DOL 8, all infants Secondary outcomes: Incidence MEF group and 26% in sample size of 191 infants per
received bolus feedings of NEC, sepsis, mortality, the NPO group (P = .34); group would be needed for an
(20 mL/kg/d) and were intraventricular hemorrhage, length infants who expired were appropriately powered study.
followed until 1 week after of hospital stay excluded from the analysis.
achievement of full enteral
nutrition (150 mL/kg/d).
Single-center NICU (Illinois,
USA)
N = 84 (n = 41 MEF group, n
= 43 NPO group)
(continued)
7. Table 2. (continued)
Study Design,
Author, Year Quality Population, Setting, N Study Objective Results Comments
Caple,30 2004 RCT Infants ≤35 week GA To determine whether infants fed Incidence of NEC: 4.2% Advancement of feeds at 30 mL/
Nonblinded with BW 1000–2000 g, initially and advanced at 30 mL/ (3/72) in the rapid- kg/d is as safe as 20 mL/kg/d with
Rate of attrition: admitted between 1994 and kg/d achieve full enteral nutrition advancement group and a comparable incidence of NEC
5/160 1995, and followed until sooner than infants fed initially and 2.4% (2/83) in the slow- between groups.
Intention-to- hospital discharge or the advanced at 20 mL/kg/d (goal 150 advancement group; P No information on mortality reported
treat analysis development of NEC mL/kg/d) value not given Authors recommend a large,
Single-center level II and III Primary outcome: Time to 3.2% overall incidence of multicenter prospective trial to
NICU at community-based achievement of full enteral NEC (RR, 1.73; 95% evaluate ways of optimizing
county hospital (Texas, nutrition CI, 0.30–10.06; P = .66) enteral nutrition without increasing
USA) Secondary outcomes: Incidence of for infants enrolled in morbidity.
N = 155 (n = 72 rapid group, NEC (Bell’s stage ≥II) and feeding study; 4.1% in preterm
n = 83 slow group) complications, length of hospital infants (1000–2000 g)
stay, duration of intravenous fluid, during the same period
weight gain not enrolled in the study
van Elburg,28 RCT Preterm infants <37 weeks To evaluate the effect of MEF (12 × Incidence of NEC: 0% MEF of preterm infants with IUGR
2004 Nonblinded GA with BW <2000 g 0.5 mL daily if BW <1000 g or 12 (0/20) in the MEF and had no effect on the development
Rate of attrition: and BW for GA <10th × 1 mL daily if BW >1000 g) on 4.5% (1/22) in the NPO of NEC.
14/56 percentile (IUGR), admitted intestinal permeability and feeding group (P = .76) No significant difference in mortality
from 1/1998 to 11/2000, tolerance between groups
enrolled within 48 hours Primary outcome: Functional However, a larger sample size
of birth and followed for integrity of the small bowel is needed to draw definitive
5 days Secondary outcomes: Feeding conclusions on the effect of MEF
Single-center NICU in tolerance, growth, and incidence of on measures of clinical outcome.
tertiary care referral center NEC (Bell’s stage ≥II)
Downloaded from pen.sagepub.com by guest on August 24, 2012
(Netherlands)
N = 56 (n = 28 MEF group, n
= 28 NPO group)
Salhotra,31 2004 RCT Infants with BW <1250 g To evaluate the tolerance of rapid (30 Incidence of NEC: 7.4% Stable VLBW infants appear to
Nonblinded subject to gastrointestinal mL/kg/d) vs slow (15 mL/kg/d) (2/27) in the rapid- tolerate rapid advancements
Rate of attrition: priming (5 mL/kg/d) via advancement of enteral nutrition to advancement group and of enteral nutrition without an
19/53 intermittent nasogastric goal of 180 mL/kg/d 0% (0/26) in the slow- increased risk of NEC.
tube bolus, on DOL Primary outcome: Time to achieve advancement group; no There was no significant difference
1–2, then randomized at 48 full enteral nutrition P value given in mortality between groups.
hours of life Secondary outcomes: Incidence of NEC-related mortality: Of note, of infants randomized to
Tertiary-level teaching NEC (Bell’s stage ≥II) and apnea There were 2 cases of the fast group, 74% completed the
hospital (India) NEC (DOL 6 and 8), trial vs 53.8% in the slow group.
N = 53 (n = 27 fast group, n = both in the fast group, The small sample size precludes
7
26 slow group) and those infants any firm conclusion on the risk of
died with associated NEC.
septicemia.
(continued)
8. Table 2. (continued)
Study Design,
8
Author, Year Quality Population, Setting, N Study Objective Results Comments
Berseth,26 2003 RCT Infants <32 weeks appropriate To compare the risks and benefits Incidence of NEC: 10% Higher risk for NEC in preterm
Nonblinded for GA, with feeds begun of enteral nutrition advancement (7/70) in the advancing infants when fed advancing
Rate of attrition: at the discretion of the (20 mL/kg/d to goal 140 mL/kg/d) group and 1.4% (1/71) feeding volumes compared with
3/144 neonatologist, and admitted compared with stable feeding in the control group (P low/stable feeding volumes over a
between 1/1996 and 1/2000 volumes (20 mL/kg/d) over a 10- = .03) 10-day period
Single-center NICU (Texas, day period The study was closed Mortality was similar in both groups
USA) Primary outcome: Incidence of NEC early due to the high (4.2% vs 4.3%, P = .97).
N = 141 (n = 70 advancing (Bell’s stage ≥II) incidence of NEC in the Mean age of enteral nutrition
group, n = 71 control Secondary outcomes: Maturation advancing group. initiation was 13 days for the 7
group) of intestinal motor patterns, time infants in the advancing group
to reach full enteral nutrition, who developed NEC. The 1
incidence of late sepsis infant in the control group who
developed NEC was fed at age 4
days. The age at feed initiation
was only given for infants who
developed NEC.
Authors conclude that minimal
feeding volumes should be
evaluated until future trials further
assess the safety of advancing
feeding volumes.
BW, birth weight; CI, confidence interval; DOL, day of life; ELBW, extremely low birth weight; GA, gestational age; IUGR, intrauterine growth restriction; MEF, minimal enteral feeding; NEC, necrotizing
enterocolitis; NICU, neonatal intensive care unit; NPO, nil per os; RCT, randomized controlled trial; RR, relative risk; VLBW, very low birth weight.
Downloaded from pen.sagepub.com by guest on August 24, 2012
Table 3. GRADE Table Question 1: When and how should feeds be started in infants at high risk for NEC?
Overall
GRADE of Evidence for Recommendation
Comparison Outcome Quantity, Type Evidence Findings Outcome GRADE
Rapid vs slow enteral nutrition advancement29-31 Incidence of NEC 3 RCTs No difference Low Weak
25
Early vs delayed minimal enteral feeding Incidence of NEC 1 RCT No difference Low Weak
27,28
Minimal enteral feeding vs nil per os Incidence of NEC 2 RCTs No difference Low Weak
Advancing vs low/stable feeding volume26 Incidence of NEC 1 RCT Higher Low Weak
NEC, necrotizing enterocolitis; RCT, randomized controlled trial.
9. Table 4. Evidence Table Question 2: Does the provision of mother’s milk reduce the risk of developing NEC relative to bovine-based products or formula?
Author, Year Study Design, Quality Population, Setting, N Study Objective Results Comments
32
Sullivan, RCT Premature infants with BW To evaluate the health benefits of Incidence of NEC: 7% (5/71) in the The rates of NEC and NEC
2010 Nonblinded 500–1250 g, fed HM (MM an exclusively HM-based diet HM40 group, 4.5% (3/67) in the requiring surgery were
Rate of attrition: 31/207 and/or DM) within the (MM and/or DM) compared HM100 group, and 16% (11/69) in markedly lower in the
first 21 days after birth, with a diet of both human and the BOV group (P = .05 between groups fed exclusively
followed until 91 days old, BOV-based products 3 groups) HM (MM and/or DM)
hospital discharge, or the Primary outcome: Fewer cases of NEC in the HM40 compared with BOV-
achievement of 50% oral PN duration and HM100 groups, with P = .09 based products.
feeds (goal 160 mL/kg/d) Secondary outcomes: Incidence between HM40 and BOV groups, 50% reduction in the
Multicenter—12 NICUs (11 of NEC (Bell’s stage ≥II), late- P = 0.04 between HM100 and incidence of NEC and
USA; 1 Austria) onset sepsis, growth, morbidity BOV, and P = .02 between HM almost 90% reduction in
Total N = 207 (n = 71, 40 (40+100) and BOV groups the incidence of surgical
mL/kg/d, HM40 group; For NEC cases requiring surgical NEC in infants fed
(HM40 mL/kg/d); intervention, P = .03 between exclusive HM (MM and/
n = 67, 100 mL/kg/d, the HM40 and HM100 groups or DM) vs BOV-based
HM100 group; (HM100 independently compared with the products
mL/kg/d); n = 69, bovine BOV group and P = .007 for HM Using exclusively HM-based
milk, BOV group) (100+40) compared with the BOV diet, NNT to prevent 1
group case of NEC is 10 and to
Exclusive HM diet (OR 0.23, 95% prevent 1 case of surgical
CI 0.08-0.66, P=0.007) NEC or death is 8.
Schanler,34 RCT Premature infants <30 weeks To determine the incidence of NEC Incidence of NEC: 6% (4/70) in the As a substitute for MM,
2005 Blinded to group GA, stratified by GA and in infants receiving pasteurized MM group, 6% (5/78) in the DM pasteurized DM offered
assignment receipt of prenatal steroids, DM vs PF as a substitute for group, and 11% (10/88) in the PF no observed short-term
(caregivers) admitted between 8/1997 MM to achieve goal 160 mL/ group (P = .39 between MM and advantage over PF for
Rate of attrition: 8/243 and 7/2001, and followed kg/d DM+PF and P = .27 between DM feeding premature infants.
Downloaded from pen.sagepub.com by guest on August 24, 2012
from birth to 90 days of Primary outcomes: and PF) There was no significant
age or hospital discharge Incidence of NEC (Bell’s stage difference in mortality
Nurseries at Texas Children’s ≥II) and late-onset sepsis between the groups.
Hospital (Texas) Secondary outcomes: Duration
N = 243 (n = 70 MM only, n of hospitalization, growth,
= 81 DM, n = 92 PF) mortality
Sisk,35 2007 OBS Infants with BW 700–1500 To determine if a high proportion Incidence of NEC: 3.2% (5/156) in Enteral nutrition containing
Prospective cohort study g, born from 5/2001 to of MM (HMM, ≥50% of enteral the HMM group and 10.6% (5/46) ≥50% MM within the
Rate of attrition: 3/202 8/2003, and followed nutrition) protects against in the LMM group (OR, 0.17; first 14 days after birth
Analysis of covariance during DOL 1–14. Infants the development of NEC as 95% CI, 0.04–0.68; P = .01 after is associated with a
and logistic regression were started on PN 1–2 compared with a low proportion adjustment for GA) significant (6-fold)
analysis days after birth if GA <30 of MM (LMM, <50% of enteral The overall incidence of NEC decreased risk of NEC.
weeks. Goal feeds were nutrition) within the first 14 days negatively correlated with the No difference in the
100–120 mL/kg/d. of life proportion of MM fed in the first incidence of surgical NEC
14 days of life (OR, 0.62; CI, or mortality between
9
0.51–0.77; P = .02) after adjustment groups
for GA.
(continued)
10. 10
Table 4. (continued)
Author, Year Study Design, Quality Population, Setting, N Study Objective Results Comments
Study supported Level III obstetric referral Primary outcome: For every 25% increase in MM
by International center for women (North Incidence of NEC (Bell’s stage proportion, the odds of NEC
Lactation Consultant Carolina, USA) ≥II) decreased by 38%.
Association, N = 202 (n = 156 HMM, n = Secondary outcomes: Feeding
University of 46 LMM) tolerance, late-onset sepsis,
North Carolina chronic lung disease, retinopathy
(Greensboro), of prematurity
and Wake Forest
University School of
Medicine
Furman,36 OBS Infants <33 weeks GA with To evaluate the dose effect of MM Incidence of NEC: 8% (3/40) group The amount of MM
2003 Prospective BW 600–1499 g, admitted compared with PF on neonatal I (used as basis of comparison administered does not
Rate of attrition: not between 1/1997 and morbidity against groups II–IV), 7% (2/29) affect the incidence of
specified 2/1999, followed through Primary outcome: Neonatal group II (OR, 1.15; 95% CI, NEC in VLBW infants.
Regression analysis week 4 of life morbidity 0.8–12.13), 11% (2/18) group III
Supported by NIH Urban tertiary care NICU Secondary outcomes: Incidence (OR, 1.99; CI, 0.14–21.03), 0%
grant M0100080 and (Ohio, USA) of NEC (Bell’s stage ≥II), (0/32) group IV 0/32 (OR, 0; CI,
University Hospitals N = 119 (n = 40 group I: sepsis, length of hospital stay 0–3.56)
of Cleveland Ohio PF only; n = 29 group II: and ventilator dependence,
MM 1–24 mL/kg/d; n = 18 retinopathy of prematurity,
Group III: MM 25–49 mL/ chronic lung disease
kg; n = 32 group IV: MM
≥50 mL/kg/d)
Downloaded from pen.sagepub.com by guest on August 24, 2012
Lambert,33 OBS Infants >36 weeks GA, born To determine possible explanations Incidence of NEC: Overall incidence Formula feeding compared
2007 Retrospective from 2/2001 to 6/2006 for why patients develop NEC of 0.51% (30/5877) with exclusive MM
Rate of attrition: 12/30 Intermountain Healthcare by comparison to age-matched Patients with NEC more likely feeding is one factor that
Descriptive statistics NICUs (Utah, USA) patients without NEC exclusively fed with PF (53%, may predispose infants to
N = 5877 (n = 30 with NEC) Primary outcome: Incidence of 16/30) or PF+MM (43%, 13/30) the development of NEC.
NEC (Bell’s stage ≥II) vs exclusive MM (3%, 1/30); no
Secondary outcomes: Morbidity, P value
feeding tolerance, length of NEC-related mortality: 13%
hospital stay mortality in infants with NEC
BOV, bovine milk; BW, birth weight; CI, confidence interval; DM, donor milk; DOL, day of life; GA, gestational age; HMM, high mother’s milk; HM, human milk; LMM, low mother’s milk; MM,
mother’s milk; NEC, necrotizing enterocolitis; NICU, neonatal intensive care unit; NIH, National Institutes of Health; NNT, number needed to treat; OBS, observational study; OR, odds ratio; PF, preterm
formula; PN, parenteral nutrition; RCT, randomized controlled trial; VLBW, very low birth weight.
11. Table 5. GRADE Table Question 2: Does the provision of mother’s milk reduce the risk of developing NEC relative to bovine-based products or formula?
Quantity, Overall
Type GRADE of Evidence Recommendation
Comparison Outcome Evidence Findings for Outcome GRADE
Human milk (mother’s or donor) vs human milk (mother’s or Incidence of NEC 1 RCT Lower Low Weak
donor) + bovine-based products32
Donor milk vs preterm formula34 Incidence of NEC 1 RCT No difference Moderate Weak
33
Mother’s milk vs preterm formula Incidence of NEC 1 OBS Lower Low Weak
Mother’s milk ≥50% vs <50% from days of life 1–1435 Incidence of NEC 1 OBS Lower Low Weak
36
Mother’s milk high vs low dose (mL/kg/d) Incidence of NEC 1 OBS No difference Low Weak
NEC, necrotizing enterocolitis; OBS, observational study; RCT, randomized controlled trial.
Table 6. Evidence Table Question 3: Do probiotics reduce the risk of developing NEC?
Author, Study Design,
Year Quality Population, Setting, N Study Objective Results Comments
37
Braga, RCT Preterm infants with BW To assess whether oral Incidence of NEC: 0% (0/119) in Probiotic use decreased the
2011 Double blind 750–1499 g, with no congenital supplementation of Lactobacillus the probiotic group and 3.6% incidence of NEC.
Rate of attrition: infections, fed MM or DM, and casei and Bifidobacterium breve (4/112) in the control group (P No difference in mortality between
62/243 admitted 5/2007 to 4/2008 prevents the occurrence of NEC = .05) groups
Randomized and started treatment Primary outcome: NEC (Bell’s stage Infants in the probiotic group
on DOL 2, duration of ≥II) reached full enteral nutrition faster
Downloaded from pen.sagepub.com by guest on August 24, 2012
treatment until DOL 30, NEC than the control group (P = .02),
diagnosis, hospital discharge, suggesting that probiotics may
or death play a role in intestinal motility.
Single-center NICU (Brazil) The study was interrupted by an
N = 231 (n = 119, probiotic external committee after 1 year,
group; n = 112, control group) due to the significant differences
found between groups.
Awad,38 RCT Neonates 28–41 weeks GA with To evaluate the use of the probiotic Incidence of NEC: Decreased incidence of NEC in
2010 Placebo control BW 1100–4300 g, with normal Lactobacillus acidophilus in the 1.7% (1/60) in the LP group, 1.7% infants who received probiotics
Double blind C-reactive protein and negative prevention of neonatal sepsis and (1/60) in the KP group, and vs placebo. However, there was
Rate of attrition: blood cultures, admitted from NEC and to further investigate 16.7% (5/30) in the placebo no difference in the incidence of
25/150 1/2006 to 5/2007 on DOL 1 and differences between LP and KP in group; P value not provided NEC between LP and KP groups.
followed until hospital discharge comparison to placebo LP group (OR 0.53, 95% CI 0.16- LP and KP were preventative
Single-center NICU (Egypt) Primary outcomes: NEC (Bell’s 1.74), KP group (OR 0.085; 95% against NEC, whereas placebo
11
N = 150 (n = 60 LP, n = 60 KP, n stage ≥II), sepsis CI 0.009-0.76), and placebo group was a risk factor for the
= 30 placebo) (OR 3.4, 95% CI 2.44-4.72) development of NEC.
(continued)
12. Table 6. (continued)
12
Author, Study Design,
Year Quality Population, Setting, N Study Objective Results Comments
Mihatsch,43 RCT VLBW infants <30 weeks GA To investigate whether Incidence of NEC: 2.2% (2/91) in No significant effect of B lactis on
2010 Double-blinded with BW <1500 g, fed PF supplementation with the probiotic the probiotic group and 4.5% the incidence of NEC
Placebo controlled or MM, admitted 5/2000 to Bifidobacterium lactis for a (4/89) in the control group (P These study results emphasize the
Rate of attrition: 8/2003 duration of 6 weeks reduces the = NS importance of an adequately
20/183 Single-center NICU, Children’s incidence of nosocomial infection NEC-related mortality: 1.1% powered trial to evaluate NEC as
Intention-to-treat Hospital (Germany) Primary outcome: Nosocomial (1/91) in the probiotic group and the primary outcome.
analysis N = 180 (n = 91 probiotic group, infection 0% (0/89) in the control group,
n = 89 control group) Secondary outcome: NEC (Bell’s P = NS
stage ≥II)
Samanta,40 RCT VLBW infants <32 weeks GA To evaluate the effect of probiotics NEC incidence: 5.5% (5/91) in the Probiotics given to VLBW infants
2009 Double-blinded with BW <1500 g, having (mixture of Bifidobacteria bifidum, probiotic group and 15.8% (15/95) reduce the incidence of NEC.
Rate of attrition: survived beyond DOL 2, fed Bifidobacteria infantis, Bifidobac- in the control group (P = .042)
18/186 exclusively MM, and admitted teria longum, and Lactobacillus The severity of NEC was similar in
between 10/2007 and 3/2008 acidophilus) on feeding tolerance, both groups (P = .62).
the incidence/severity of NEC Mortality due to NEC or sepsis
Single-center NICU (India) (Bell’s stage ≥II), and mortality was lower in the probiotics
N = 186 (n = 91 probiotic group, related to NEC or sepsis group compared with the
n = 95 control group) Primary outcomes: Feeding control group (4.4% vs 14.7%,
tolerance, length of hospitalization, respectively; P = .032).
morbidities (NEC, sepsis, and/or
death due to NEC or sepsis)
Lin,39 2008 RCT Infants <34 weeks GA with To investigate the efficacy of oral Incidence of NEC: 1.8% (4/217) The administration of probiotics to
Investigators at each BW <1500 g, fed MM or PF, probiotics, Bifidobacterium in the probiotic group and 6.5% preterm VLBW infants reduced
Downloaded from pen.sagepub.com by guest on August 24, 2012
center and the admitted between 4/2005 and bifidum and Lactobacillus (14/217) in the control group (P the incidence of NEC but did not
breast milk team 5/2007 acidophilus (Infloran); 125 mg/kg/ = .02) affect NEC-related mortality.
not blinded but Seven level III NICUs (Taiwan) dose twice daily for 6 weeks in the NEC-related mortality: 0.9% Based on study results, NNT to
were not involved N = 434 (n = 217 probiotic group, prevention of NEC (2/217) in the probiotic group prevent 1 case of NEC was 20
in the care of the n = 217 control group) Primary outcomes: NEC (Bell’s and 1.4% (3/217) in the control patients.
study infants stage ≥II) and death group (P = .98)
Multicenter Secondary outcomes: Culture-proven
Rate of attrition: sepsis without NEC, chronic
20/443 lung disease, periventricular
leukomalacia, intraventricular
hemorrhage, feeding amount and
weight gain per week, days to full
enteral nutrition
(continued)
