2. It is recommended that donor milk be considered, if available and affordable, as an alternative, when mother's own milk is unavailable (Schanler et al., 2005 [A]; Local Expert Consensus [E]).

   Note 1: Current cost of donor milk is 4 to 5 times higher than formula designed for premature infants.

   Note 2: When recommending the use of donor milk, reimbursement issues may need to be considered and discussed with the family. Donor milk may not be covered by insurance. For more information see www.hmbana.org.

Feeding Strategies

Available studies were reviewed for use of minimal enteral feeds, timing of initiation of feeds, and rates for advancing feeds (see Table below).

Table: Definitions Used in Studies of Feeding Issues

Minimal Enteral Feeding

3. There is insufficient evidence regarding the role of minimal enteral feedings in preventing NEC.

   Note 1: A meta-analysis reviewing 6 studies showed use of minimal feeds had no effect on risk of NEC (Tyson & Kennedy, 2003 [M]).

   Note 2: Inclusion of two additional randomized controlled trials (RCTs) did not alter results of the meta-analysis (McClure & Newell, 2000 [A]; Schanler et al., 1999 [A]).

Timing of Initiation of Feeds

4. There is insufficient evidence to support either early or delayed initiation of feeding relative to risk for NEC (Kennedy, Tyson, & Chamnanvanikij, "Early versus delayed," 2003 [M]; Wilson et al., 1997 [A]; LaGamma, Ostertag, & Birenbaum, 1985 [C]; McKeown et al., 1992 [D]).

Rate of Advancement of Feeds

5. There is insufficient evidence to recommend a specific rate of feeding volume advancement in relation to NEC risk.

   Note 1: No difference in NEC risk was observed in studies advancement as low as 10 cc/kg/day and as high as 35 cc/kg/day (Kennedy, Tyson, & Chamnanvanikij, "Rapid versus slow," 2003 [M]; Kamitsuka, Horton, & Williams, 2000 [D]).

   Note 2: One large study showed a decreased rate of NEC in infants maintained at 20 cc compared to those children advanced. Inclusion of this study did not alter results of the meta-analysis (Berseth, Bisquera, & Paje, 2003 [A]).

Transpyloric Versus Gastric Feeding

6. There is insufficient evidence to support either transpyloric or gastric feeding methods relative to the risk of NEC (McGuire & Anthony, "Formula milk versus term human milk," 2003 [M]).

Bolus Versus Continuous Feeding

7. There is insufficient evidence to support either bolus or continuous tube feeding as a method to reduce the risk of NEC (Premji & Chessell, 2003 [M]).

Umbilical Artery Catheters (UAC)

8. There is insufficient evidence to recommend a specific placement location for the tip of the umbilical artery catheter. Umbilical artery catheter position (high versus low) has not been found to affect the incidence of NEC (Barrington, 2003 [M]).
9. There is insufficient evidence to evaluate the risk of NEC associated with enteral feeding while an umbilical artery catheter is in place.

   Note: One small randomized trial found no difference in the incidence of NEC between infants fed early, with a UAC in place, and those in which feeds were delayed until 24 hours after UAC removal (Davey et al., 1994 [B]).

Probiotics

10. There is insufficient evidence to recommend either use or avoidance of probiotics. Clinical trials of the effects of probiotics on the risk of NEC have not consistently shown benefit.

    Note 1: One large observational study showed a significant decrease in NEC compared to historical controls when infants were treated with lactobacillus and bifidobacterium (Hoyos, 1999 [D]).

    Note 2: One large RCT of lactobacillus GG found no significant effect on the incidence of NEC in very low birth weight (VLBW) infants (Dani et al., 2002 [A]).

    Note 3: A second large RCT of lactobacillus and bifidobacterium in VLBW infants found a significant reduction in NEC incidence among treated infants (Lin et al., 2005 [A]).

    Note 4: Issues regarding the safety of probiotic use in immunodeficient hosts such as preterm infants have been raised, particularly the potential for sepsis. None of the clinical trials reported harmful side effects although they were not powered to evaluate safety (Land et al., 2005 [O]).

Additional Prevention Strategies

11. The following prevention strategies have been evaluated in large sample size, randomized, controlled trials or meta-analyses with no statistically significant effects on NEC risk.

    Amino Acid Supplementation

    • Enteral glutamine (Vaughn et al., 2003 [A])
    • Intravenous (IV) glutamine (Poindexter et al., 2004 [A])
    • Enteral arginine (Amin et al., 2002 [A])

    Immunoglobulins

    • Oral immunoglobulin G (IgG) or IgA/IgG combination (Foster & Cole, 2004 [M]; Rubaltelli, Benini, & Sala, 1991 [A]; Eibl et al., 1988 [A])
      Note: There is no RCT of IgA alone for the prevention of NEC.
    • Intravenous IgG (Fanaroff et al., 1994 [A])

    Antibiotics

    • Results from a meta-analysis of 5 studies suggest that oral aminoglycosides decrease the incidence of NEC. However, lack of information on other outcomes including mortality and development of resistant bacteria precludes any recommendation (Bury & Tudehope, 2003 [M]).
    • Oral or intravenous erythromycin (Ng & Shah, 2003 [M])

Where evidence is present but sparse, information is shared below for purposes of awareness even if conclusions can not be made.

Acidification

12. It is recommended that histamine-2 receptor (H2) blocker therapy for gastric acidity be used with caution. Two studies suggest that gastric pH alters risk for NEC.

    Note 1: One small RCT showed that acidification of feeds was associated with decreased risk of NEC (Carrion & Egan, 1990 [B]).

    Note 2: One large retrospective case-control study showed a significant association between H2-blocker therapy and higher rates of NEC (Guillet et al., 2006 [D]).

Supplemental Vitamin E

13. In those at greatest risk for NEC, studies suggest that supplemental vitamin E may increase the risk of NEC with no effect on other outcomes with the exception of risk of severe retinopathy of prematurity which may be less (Brion, Bell, & Raghuveer, 2003 [M]; Johnson et al., 1985 [A]; Finer et al., 1984 [D]).

Prenatal Indomethacin

14. There is insufficient evidence to recommend either use or avoidance of prenatal indomethacin related to risk of NEC (Parilla et al., 2000 [D]; Vermillion & Newman, 1999 [D]; Major et al., 1994 [D]; Norton et al., 1993 [D]).

Postnatal Indomethacin

15. Evidence does not support an altered risk of NEC with use of indomethacin for prevention of intraventricular hemorrhage (Fowlie & Davis, 2003 [M]) or treatment of patent ductus arteriosus (PDA) (Malviya, Ohlsson, & Shah, 2003 [M]; Gersony et al., 1983 [A]; Cooke & Embleton, 2000 [S]).

    Note 1: Although the number of infants <1,000 gm birth weight was small, a collaborative trial showed no difference on NEC risk using surgical versus indomethacin treatment for PDA (Gersony et al., 1983 [A]).

    Note 2: No difference in NEC risk has been noted between ibuprofen and indomethacin used for PDA treatment or prevention (Shah & Ohlsson, 2003 [M]).

Medical Management

Evaluation

Clinical Assessment and Diagnosis

16. Bell's NEC staging system is commonly used to assess NEC severity (Bell et al., 1978 [C]). Studies to firmly establish assessment and diagnosis criteria are not available (see Table below).

Table: Bell's Staging for NEC

Diagnostics

Laboratory Studies

17. While specific pathogens have been isolated from stools and abdominal fluid during outbreaks of NEC, no specific pathogen has been found to have a consistent causal relationship with NEC (de la Cochetiere et al., 2004 [C]; Peter et al., 1999 [C]; Millar et al., 1996 [C]; Rotbart et al., 1988 [C]; Blakey et al., 1985 [C]; Thomas et al., 1984 [C]; Gupta, Burke, & Herson, 1994 [D]; Keller et al., 1991 [D]; Sherertz & Sarubbi, 1982 [D]).
18. There is insufficient evidence to support the use of stool patterns, presence of occult blood, or presence of specific pathogens as clinical indicators of NEC risk (Peter et al., 1999 [C]; Abramo et al., 1988 [C]; Andrews & Krowchuk, 1997 [D]).
19. There is insufficient evidence for use of gastric residuals as a predictor of NEC. Gastric residuals in infants who developed NEC tended to be larger, but significant overlap in the amount of residual precludes its use as a marker for NEC (Cobb, Carlo, & Ambalavanan, 2004 [D]).

Radiologic Studies

20. It is recommended that an abdominal radiograph be performed in infants with clinical suspicion of NEC (Bell et al., 1978 [C]). The influences on infant outcome and diagnostic validity of the number of abdominal x-rays, the type of view(s) or the frequency or timing of abdominal radiographs have not been systematically studied.

    Note 1: Inter-observer reliability of radiographic signs of NEC is low (Napoli et al., 2004 [D]).

    Note 2: While positive radiographic findings have good predictive value, negative studies must be interpreted with caution (Tam, Camberos, & Applebaum, 2002 [D]; Kosloske, 1994 [D]).

Intervention

Medications

21. There is insufficient evidence on benefit or risk regarding choice of antibiotic regimens or duration of antibiotic treatment of NEC (Faix, Polley, & Grasela, 1988 [B]; Scheifele et al., 1987 [C]).

    Note: Decisions regarding antibiotic choice and duration might best be guided by:

    • Culture results
    • Antibiotic resistance patterns present within nurseries

    (Local Expert Consensus [E])

22. There is insufficient evidence on benefit or risk of routine clindamycin use for treatment of NEC (Faix, Polley, & Grasela, 1988 [B]).

    Note: One small randomized controlled trial using clindamycin showed an increase in bowel strictures (Faix, Polley, & Grasela, 1988 [B]).

23. There is insufficient evidence on benefit or risk of oral aminoglycoside use for the treatment of NEC (Hansen et al., 1980 [B]).

Nutrition

24. There is insufficient evidence on benefit or risk regarding the timing of reinitiating feeding once the diagnosis of NEC has been made.

    Note: One retrospective study evaluating the impact of early initiation of feeding (<10 days from diagnosis) suggests a positive impact. Early feeding was associated with a shorter time to full feeds, less catheter related sepsis, and a shorter hospital stay. The study was underpowered to evaluate recurrence risk of NEC (Bohnhorst et al., 2003 [C]).

Paracentesis

25. Abdominal paracentesis may be helpful to confirm the presence of intestinal gangrene in infants with NEC (Ricketts, 1986 [D]; Kosloske & Goldthorn, 1982 [D]). Indications for paracentesis are absence of pneumoperitoneum and one of the following:
    • Portal venous gas
    • Erythema of abdominal wall
    • Fixed, tender abdominal mass
    • Persistently dilated intestinal segment
    • Clinical deterioration (see Table "Relative indications for surgical consultation," below)

    Note: Positive results reliably predict the presence of intestinal gangrene (accuracy 90% to 97.5%). A "positive" result is considered an aspiration of 0.5 mL of peritoneal fluid and one of the following:

    • Yellow-brown or brown staining fluid
    • Gram stain positive for bacteria

    However, a negative result does not reliably exclude the presence of intestinal gangrene (40% false negative rate) (Kosloske, 1994 [D]; Ricketts, 1986 [D]).

Surgical Management

Site of Care

26. It is recommended that neonates with suspected NEC be transferred to a facility that can meet their possible surgical needs if surgical services are not available within the admitting institution. A safe transfer is best achieved when the child is hemodynamically stable. Indications for transfer might include:
    • Pneumoperitoneum
    • Radiographic evidence of portal venous gas (Buras et al., 1986 [D])

    (Local Expert Consensus [E]).

    Note: Evidence suggests that portal gas or diffuse pneumatosis is associated with more severe disease (see Evaluation and Intervention Recommendation #28 below).

Evaluation and Intervention

27. It is recommended that neonates with clinical/radiographic evidence of pneumoperitoneum be evaluated by a surgeon in a center in which operative intervention can be performed if indicated. Pneumoperitoneum represents an absolute indication for surgical intervention (Tam, Camberos, & Applebaum, 2002 [D]; Kosloske, 1994 [D]; Kosloske, Papile, & Burstein, 1980 [D]).
28. It is recommended that neonates with clinical/radiographic evidence of intrahepatic venous gas (portal venous gas) be evaluated by a surgeon in a center in which operative intervention can be performed if indicated (Molik et al., 2001 [C]; Tam, Camberos, & Applebaum, 2002 [D]; Kosloske, 1994 [D]; Rowe et al., 1994 [D]; Kennedy, Holt, & Ricketts, 1987 [D]; Buras et al., 1986 [D]).
29. The following conditions may be considered as relative indications for surgical consultation in a center in which operative intervention can be performed if indicated:
    • Abdominal wall cellulitis
    • Fixed dilated intestinal segment by x-ray
    • Tender abdominal mass
    • Clinical deterioration refractory to medical management (see Table below)

    (Local Expert Consensus [E])

Table 3: Relative Indications for Surgical Consultation

(Ververidis et al., 2001 [D]; Gupta, Burke, & Herson, 1994 [D]; Buras et al., 1986 [D]; Local Expert Consensus [E])

30. Radiographic signs of NEC have high specificity but low sensitivity, with poor negative predictive values. It is recommended that the decisions regarding surgical intervention be based on both the clinical and radiological presentation. It is recommended that decisions regarding the need for surgical intervention not be made solely on the basis of absent radiographic signs (Tam, Camberos, & Applebaum, 2002 [D]; Kosloske, 1994 [D]).
31. One retrospective study compared the use of the neonatal intensive care unit for surgery versus transport to an operating room setting for neonates weighing less than 1,500 g with severe NEC. Transport of neonates less than 1,500 g from an intensive care unit to an operating room is associated with deterioration in physiologic parameters (Frawley, Bayley, & Chondros, 1999 [D]).
32. The role of percutaneous drainage versus exploratory laparotomy is controversial (Moss et al., 2001 [M]; Ahmed, Ein, & Moore, 1998 [D]; Azarow et al., 1997 [D]; Morgan, Shochat, & Hartman, 1994 [D]; Cheu, Sukarochana, & Lloyd, 1988 [D]). A multicenter, prospective, randomized trial of percutaneous drainage compared to laparotomy, in 117 very low birth weight newborns with NEC and intestinal perforation, demonstrated no difference in short term outcomes (Moss et al., 2006 [A]).  No recommendation can be made on surgical approach at the time of publication of this guideline.

Definitions:

Evidence Based Classifications:

M: Meta-analysis or Systematic Review
A: Randomized controlled trial: large sample
B: Randomized controlled trial: small sample
C: Prospective trial or large case series
D: Retrospective analysis
O: Other evidence
S: Review article
E: Expert opinion or consensus
F: Basic laboratory research
L: Legal requirement
Q: Decision analysis
X: No evidence

References open in a new window

For information about availability, see the "Availability of Companion Documents" and "Patient Resources" fields below.

Electronic copies: Available in Portable Document Format (PDF) from the Cincinnati Children's Hospital Medical Center Web site.

Print copies: For information regarding the full-text guideline, print copies, or evidence-based practice support services contact the Children's Hospital Medical Center Health Policy and Clinical Effectiveness Department at HPCEInfo@chmcc.org.

• Necrotizing enterocolitis. Cincinnati Children's Hospital Medical Center, 2005 Oct. Electronic copies: Available from the Cincinnati Children's Hospital Medical Center Web site.

Please note: This patient information is intended to provide health professionals with information to share with their patients to help them better understand their health and their diagnosed disorders. By providing access to this patient information, it is not the intention of NGC to provide specific medical advice for particular patients. Rather we urge patients and their representatives to review this material and then to consult with a licensed health professional for evaluation of treatment options suitable for them as well as for diagnosis and answers to their personal medical questions. This patient information has been derived and prepared from a guideline for health care professionals included on NGC by the authors or publishers of that original guideline. The patient information is not reviewed by NGC to establish whether or not it accurately reflects the original guideline's content.

Copies of Cincinnati Children´s Hospital Medical Center (CCHMC) Evidence-Based Clinical Practice Guidelines (EBCG) are available online and may be distributed by any organization for the global purpose of improving child health outcomes. Examples of approved uses of CCHMC´s EBCG include the following:

• Copies may be provided to anyone involved in the organization's process for developing and implementing evidence-based care guidelines.
• Hyperlinks to the CCHMC website may be placed on the organization's website.
• The EBCG may be adopted or adapted for use within the organization, provided that CCHMC receives appropriate attribution on all written or electronic documents.
• Copies may be provided to patients and the clinicians who manage their care.

Notification of CCHMC at HPCEInfo@cchmc.org for any EBCG adopted, adapted, implemented or hyperlinked to by a given organization and/or user, is appreciated.

Guidelines represented on the NGC Web site are submitted by guideline developers, and are screened solely to determine that they meet the NGC Inclusion Criteria which may be found at http://www.guideline.gov/about/inclusion.aspx .

NGC, AHRQ, and its contractor ECRI Institute make no warranties concerning the content or clinical efficacy or effectiveness of the clinical practice guidelines and related materials represented on this site. Moreover, the views and opinions of developers or authors of guidelines represented on this site do not necessarily state or reflect those of NGC, AHRQ, or its contractor ECRI Institute, and inclusion or hosting of guidelines in NGC may not be used for advertising or commercial endorsement purposes.

Readers with questions regarding guideline content are directed to contact the guideline developer.