Introduction

Necrotizing enterocolitis (NEC) is the most common life-threatening gastrointestinal disease in neonates [68]. Despite three decades of research efforts, NEC remains a major cause of death for neonates undergoing surgery. The incidence of NEC has increased in the past decades, as the advantages in neonatology and the modern neonatal intensive care unit have led to the increased survival of infants of even smaller birth weight and younger gestational age [30]. Additionally, as surfactant has become standard of care in preterm infants, the number of very low birth weight (VLBW) infants at risk for developing NEC has continued to rise [29, 30, 50].

NEC occurs in one to three in 1,000 live births [37, 52], equally often in female and male [52]. NEC most commonly affects babies born between 30 and 32 weeks and is often diagnosed during the second week of life [36, 38]. The mortality from NEC has been cited as 10–50% of all affected infants [29]. The surgical mortality has decreased over the last several decades from 70% to numbers between 20 and 50% [29].

NEC has a multifactorial etiology and the pathogenesis has not fully been elucidated. The classic histological finding is coagulation necrosis present in over 90% of specimens [3]. This finding suggests the importance of ischemia in the pathogenesis of NEC [29]. Inflammation and bacterial overgrowth are also present [3]. There is an assumption that NEC occurs by the interaction of three events: Initially a mucosal injury occurs due to intestinal ischemia, followed by inflammation of the disturbed mucosal integrity with subsequent necrosis of the affected area. The further steps are colonization by pathogenic bacteria and excess protein substrate in the intestinal lumen. Furthermore the immunologic immaturity of the neonatal gut has been implicated in the development of NEC [37].

NEC affects most commonly the terminal ileum, caecum and ascending colon. Typical clinical signs include abdominal distension, bile- or blood stained emesis or gastric aspirate, abdominal wall erythema and bloody stools. Diagnosis is based on radiographic evidence as bowel distension, ileus, pneumatosis intestinalis or bowel perforation. Management includes parenteral nutrition and antibiotics, or surgical approach with bowel resection. Over the last years, different strategies for prevention of NEC have been developed. None of the strategies has been really a break through. This article presents a multi-modal approach of a prevention strategy combining well-known strategies to one concept including enteral administration of antibiotics, antifungal agent, probiotics, plus early trophic feeding with human breast milk, resulting in a low NEC incidence within a neonatal intensive care unit. Furthermore, it gives an overview of different NEC prevention strategies.

Methods

We performed a retrospective analysis of a local protocol of NEC prophylaxis. Included were all VLBW infants admitted to the neonatal intensive care unit (tertiary center), including transfers within the first 28 days of life. Two groups of infants were analyzed: patients, who were admitted on the first day of life formed the “inborn group”, all patients, who were admitted after the onset of their second day of life formed the “outborn group”. In inborn group prophylaxis of NEC according to protocol started within the first 24 h of life and continued to discharge, in outborn group prophylaxis of NEC according to protocol started after admission and was continued to discharge. Whereas inborn group represents our standard patient collective; the outborn group serves as a comparison group, having had no standardized NEC prophylaxis before admission. Being a third level universitarian neonatal intensive care unit, preterm infants have to be admitted from peripheral hospitals, if problems occurred during their stay in those units. None of those units uses a NEC-Prophylaxis protocol similar to the present.

The protocol of NEC prophylaxis consists of enteral antibiotics, enteral antifungal agent, enteral probiotics, and trophic feeding as follows: Gentamycin (7 mg/kg 12 hourly per os), Nystatin (10.000 IU/kg 6 hourly per os), and enteral probiotics (Lactobacillus Rhamnosus GG 1 g=1×109 colony forming units per day, divided in two doses).

The feeding protocol at the NICU starts with the administration of pooled donor human milk (1 ml/kg all 3 h) on the first day of life, with a stepwise increase of 1 ml/kg per day during the first week of life, with a change to expressed breast milk of the preterm infant’s mother. If no expressed breast milk of the preterm infant’s mother was available, feeding with pooled donor human milk is continued until the infant’s nutrition reaches fully enteral feeding. The aim is to reach fully enteral nutrition within the 14th day of life. Mother milk is pasteurized until the infant reaches 32 weeks of gestation or a body weight of 1,500 g.

All infants received a prophylactic administration of Indomethacin for 3 days, according to a previously published protocol [45, 46].

Various classifications have been published for NEC [5, 8, 70]. In the present study NEC was defined, using Bell’s criteria stage II or greater [5].

Results

Over a 7-year period from 1998 to 2004, 405 very-low-birth-weight infants have been admitted to the NICU, including transfers within the first 28 days of life.

Out of these 405 infants, 334 (82%) infants have been admitted within the first 24 h of life (inborn group), and 71 (18%) have been admitted after 24 h of life (outborn group). Patient’s demographical data and the incidence of NEC are shown in Table 1. There is no difference between birth weight and gestational age in the two studied groups.

Table 1 Patients’ demographical data and NEC incidence
Full size table

Five of the 347 surviving and studied infants developed clinical features of NEC. The inborn group showed a NEC incidence of 0.7% (two infants), whereas the outborn group showed a NEC incidence of 4.5% (three infants). This difference was significant (P=0.049, Fisher’s exact test).

Whereas the two NEC patients of the inborn group were admitted within the first 24 h of life, the three patients of the outborn group were admitted on 11th, 12th, and 14th day of life, respectively. Two patients of the outborn group already had diagnosis of NEC at admission and both needed surgical treatment. The third patient of the outborn group had an additionally diagnosis of phenylketonuria and therefore had to be fed differently, according to special protocol (including only small amounts of human milk). The latter patient (of outborn group) plus the two patient of the inborn group did not need a surgical intervention, they were treated with parenteral nutrition and administration of antibiotics.

Discussion

In the present study the incidence of NEC was 0.7%, when this group of patients was submitted to a standardized protocol of NEC prophylaxis, starting from the first day of life (inborn group). Shimura et al. [63] described a similar low incidence of NEC (0.6%), Lin et al. [40] described a low incidence of NEC (1.1%) in a group of patients receiving probiotics. In contrast, the incidence of NEC in Canada, USA and most other western countries reaches 5–7% [3, 27, 30, 31, 33, 60].

The infants of the outborn group cannot represent a normal control group, of course. This group was analyzed separately, not having had NEC prophylaxis according to protocol starting on first day of life. None of these infants had received NEC prophylaxis before admission. Within this outborn group incidence of NEC was according to the literature, out of this perspective the outborn group was called “control group”. Two of the three patients with NEC within this group had already been diagnosed with NEC before admission. The third patient was admitted to our NICU with an additionally diagnosis of phenylketonuria. Because of this metabolic disease the patient had to be fed according to a special protocol, allowing only small amounts of human milk. Therefore this patient was the only one, who did not receive NEC prophylaxis according to standardized protocol. The patient developed NEC after 28 days of life (during stay within our NICU).

Additionally to NEC prophylaxis protocol all infants received a prophylactic administration of indomethacin for the first 3 days to prevent intracranial hemorrhage, whereas only some of the outborn group received indomethacin. In the early 1980s the use of indomethacin has been implicated, followed by randomized controlled trials [19–21, 23] of prophylactic administration. Although these trials reported effectiveness of indomethacin administration in preventing IVH and PDA, unexpected side effects secondary to decreased splanchicus blood flow resulted in a restrain from a universal recommendation of the use of indomethacin [19]. Controversially, O’Donovan et al. [35], in a retrospective study, and Cooke et al. [13] for the Cochrane Neonatal Collaborative Review Group, concluded that prophylactic indomethacin treatment was not associated with an increased risk for the development of NEC.

Over the last years different strategies for prevention of NEC, such as changing the feeding practice, using donor breast milk, probiotics and immunoglobins have been developed. The following paragraphs summarize these different NEC prevention strategies published in the literature. However, in contrast to the multi modal approach of the present study, most of these studies used a single approach only.

Feeding practices

The GI tract is an active organ in utero. The fetus swallows amniotic fluid composed of nutrients, growth factors, and immunoglobulins [51]. Low gestational age reflects the developmental immaturity of the intestine. On the other hand, the intestine is ready to digest enteral nutrition and to tolerate bacteria and other organisms acquired after birth.

Thus, the question of fast versus slow and early versus delayed feeding has been discussed extensively in the literature, several randomized trails have shown no effect on the incidence of NEC [54] so far.

Delaying the initiation of feeds has been shown to postpone the onset of NEC. Therefore, it was a common practice to withhold feeds from premature infants especially after the initiation of parenteral nutrition. However, this practice is not without a risk: already short term starving periods in animals showed mucosal atrophy and increased permeability of the gut mucosa [57]. Although, the role of enteral feeding as a risk factor for NEC has been emphasized, 5–10% of NEC occurs in babies who have never been fed enterally [41]. Small trophic feeds have been shown to stimulate maturation of GI function, although they have not been shown to decrease the incidence of NEC [41]. Brown and Sweet [9] postulated that an aggressive enteral feeding protocol of more than 20 ml/kg/day increases the incidence of NEC. Bersteh et al. [7], randomly assigned infants to have feeding volumes increased daily by 20 ml/kg rather than being held at minimal volume for the first 10 days of life. The study stopped when seven infants in the group with advancing feeding volumes developed NEC, compared to one infant in the minimal feeding group. Thus the authors concluded, that advancing feeding volumes increased the risk of NEC.

Breast milk

Breast milk is the recommended source of enteral nutrition for preterm infants and has been demonstrated to decrease the incidence of NEC [6, 44]. Several epidemiological and animals studies indicated that breast milk is protective. Formula fed newborn infants have a six- to tenfold increase of NEC when compared with innate breast milk fed infants [41]. Therefore, it would seem prudent to consider minimal feeding volumes of breast milk, when available, rather than formula for the first 7–10 days of life. Multiple factors in breast milk are hypothesized to prevent the development of NEC, including immunoglobulins, erythropoietin, IL-10, epidermal growth factor (EGF) and platelet-activating factor (PAF)-acetylhydrolase [54].

EGF, a potent protein that produces a variety of biologic responses such as enhanced proliferation and differentiation of epithelial cells. It has been reported as an important trophic factor for the developing intestine [48]. Reduced level of salivary EGF has been identified in neonates at the time of onset of NEC compared with age-matched control neonates [64]. Breast milk, including colostrum is a major source of EGF but formula is exclusive EGF [34]. The enteral administration of recombinant EGF reduced the development and the incidence of NEC in a neonatal rat model [15].

PAF, a potent phospholipid inflammatory mediator produced by inflammatory cells, endothelial cells, platelets, and bacteria of the intestinal flora has been implicated in the pathogenesis of NEC [54]. In infants developing NEC [11, 12], elevated PAF levels and decreased levels of PAF-acetylhydrolase (PAF-AH), the enzyme responsible for the degradation of PAF, have been reported. In animal experiments using neonatal rats, an injection of PAF directly into the aorta caused an intestinal disease similar to NEC, whereas the administration of PAF-AH prevent the development of NEC in the same model [22]. Thus, the presence of PAF-AH in breast milk may contribute to its protective effect.

These findings suggest that breast milk may be the recommended source of enteral nutrition for preterm infants. The additional beneficial effect of breast milk is the delivery of immunoprotective factors to the immature gut mucosa [41].

Donor breast milk versus formula milk

When expressed breast milk of the preterm infant’s mother is not available, an alternative is banked milk from donor mothers. However, donor human milk is typically the breast milk of mothers who have delivered at term. This milk has a lower content of protein and host defence protein compared to breast milk of a mother who has delivered a preterm infant [44]. Several studies of comparison between Donor human breast milk versus formula milk in premature infants have been published [24, 25, 39, 43, 44, 53, 61, 67, 69, 71, 73]. Mc Guire and Anthony [44] summarized four small trials (Gross 1983, Lucas 1990, Svenningsen 1982 and Tyson 1983), all initiated 20 years ago. Aim of the trails was to compare the feeding of Donor human milk versus formula for preventing NEC in preterm infants. None of the trials found a statistically significant difference between human milk versus formula in regard to the incidence of NEC. Although, none of the single studies showed a significant result, the meta-analysis of all studies did show a significant reduced relative risk of NEC with feeding of Donor human milk. Furthermore, NEC was three times less likely, and confirmed NEC was four times less likely, in infants who received donor breast milk rather than formula milk [44]. However, other authors have failed to reproduce the data; they failed to show a reduced risk in the incidence of NEC by using donor human milk. Recently, Schanler et al. [61] compared preterm formula and donor human milk as substitutes for mother’s own milk in premature infants. They reported similar rates of NEC and late onset sepsis within the groups.

Probiotics

The term probiotic was derived from the Greek, meaning “for life”. An expert panel commissioned by the Food and Agriculture Organisation of the United Nations (FAO) and the World Health Organisation (WHO) defined probiotics as “Live microorganisms which when administered in adequate amounts confer a health benefit on the host.” This is the definition that should be used, and probiotics should not be referred to as biotherapeutic agents [55]. The range of effects of probiotics on the gut are wide and include changes of intestinal permeability, enhanced mucosal IgA response, increases in the production of anti-inflammatory cytokines and protection of the mucosa against colonization from pathogens [49].

The intestinal microflora in VLBW infants may be dominated by many pathogens such as Enterococcus faecalis, Escherichia coli, Staphylococcus epidermidis, Enterobacter cloacae, Klebsiella pneumonia, and Staphylococcus haemolyticus [28, 55]. Several studies have investigated the intestinal microflora of infants with NEC. Clostridium perfringens has been isolated from 40% of infants with necrotizing entercolitis, compared with 13% of controls [55]. Bell et al. [28] described increased numbers of gram-negative bacteria (in particular Escherichia coli and Klebsiella) in the stools of infants with NEC.

Based on these observation and previously published data from animal studies, it has been postulated that probiotics may offer similar protection against NEC in premature infants. Several studies have used different strains of probiotics and different administration regimes (length of treatment, dosage) in preterm infants. None of the trials has reported adverse effects, furthermore, there was not observed any episode of pathogenic infection caused by a probiotic organism [1, 4, 14, 32, 40].

A successful colonization rate of 80–90% for Lactobacillus GG has been reported in previous published studies. In contrast, Agarwal et al. [1] showed colonization about 25% in VLBW infants with Lactobacillus GG and suggested that colonization depends on the interplay of multiple factors in the intestinal milieu. For Lactobacillus acidophilus a successful colonization rate of 60–86% has been reported. Hoyos et al. [32] showed a 60% decrease of NEC in neonates during 1 year in a Columbian NICU using two strains of probiotic Lactobacillus acidophilus and Bifidobacterium infantis. They reported a threefold decrease of NEC cases and a fourfold decrease in NEC mortality; however, the comparison was with historical controls. In a prospective randomized controlled trial, Lin et al. [40] reported a significantly lower incidence of NEC in a group of preterm infants receiving probiotics (1.1%), versus the control group (5.4%).

Larger clinical trials are necessary to evaluate the safetyness and efficacy of this promising intervention, to better define both the benefits and the risks for premature infants. Recently, Schultz et al. [62] described a possible further direction for the administration of probiotics. They showed that the temporary colonization of an infant with L. GG may be possible by colonizing the pregnant mother before delivery.

Immunoglobulins

A number of reports have been published, which suggest that orally taken immunoglobulins (IgA and IgG) have an immunoprotective effect on the gastrointestinal mucosa [16, 59]. Premature infants have decreased levels of immunoglobulins, especially secretory IgA [17]. In a randomized clinical trial, Eibl et al. [16] evaluated the efficacy of an oral immunoglobulin preparation (73% IgA and 26% IgG) in reducing the incidence of NEC in infants of low birth weight for whom breast milk from their mothers was not available. They reported no cases of NEC in the treatment group of 88 infants compared with six cases of NEC in the control group of 91 infants, respectively. Rubaltelli et al. [59] evaluated in a randomized clinical trial the efficacy of an oral immunoglobulin preparation (containing monomeric IgG in a concentration of 90%) in reducing the incidence of NEC in infants of VLBW for whom maternal breast milk was not available. They reported not any case of NEC in the treatment group of 65 infants compared with four cases of NEC in the control group of 67 infants, respectively. However, other authors have failed to reproduce the data, they failed to show a decrease in the incidence of NEC by using oral immunoglobulins. In a prospective randomized trial, Fast et al. [17] compared the efficacy of oral gentamycin versus oral IgA–IgG for the prophylaxis of NEC. NEC was diagnosed in 13 cases in the oral IgA–IgG group of 100 infants compared with one case in the oral gentamycin group of 100 infants. Richter et al. [56] examined the efficacy of oral IgG prophylaxis for the prevention of NEC compared to a historical cohort group; they reported no difference in the incidence in both groups and concluded that infants were not protected against NEC by the use of oral IgG. For the Cochrane Neonatal Collaborative Review Group, Foster and Cole [18] recently concluded that based on the available trials, the evidence does not support the administration of oral immunoglobulin for the prevention of NEC.

Oral antibiotics

Published data suggest that the use of enteral antibiotics may be effective as NEC prophylaxis. Grylack and Scanlon [26] evaluated the effects of prophylactic oral gentamycin therapy in the prevention of NEC. In their study, none of the 20 gentamycin treated infants developed NEC, whereas four (of 22 infants) within the control group. In contrast, Rowley and Dahlenburg [58] reported no decrease in the incidence of NEC using an oral gentamycin regimen. Recently, in a prospective, double blind, randomized, placebo controlled study, Siu et al. [66] evaluated the effectiveness of oral vancomycin in the prophylaxis of NEC. They reported a NEC incidence of 13% (9 of 71) in the group of infants receiving oral vancomycin, compared to a NEC incidence of 28% (19 of 69) in the group receiving the placebo solution. For the Cochrane Neonatal Collaborative Review Group, Bury and Tudehope [10], evaluated five trials where oral antibiotics were used as prophylaxis against NEC in low birth weight and preterm infants. Their analysis suggests that oral administration of prophylactic enteral antibiotics results in a statistically significant reduction of NEC and in NEC-related deaths.

In a recently published article, Bell [4] summarizes different prevention strategies. Calculating the numbers needed to treat for the different strategies to prevent NEC, the most effective strategy was the administration of enteral antibiotics, followed by human breast milk feeding. However, the risks of enteral antibiotics have not been quantified yet, thus this strategy has never been widely adopted, due to concerns about the emergence of resistant bacteria and absorption of antibiotics from the gut [4]. However, such adverse effects have not been reported so far.

Oral antifungal agents

Mucocutaneus candidiasis (oral, perineal, other skin sites) is a frequent finding in the neonatal unit (3.2% of all admissions Gupta 1996, 7.8% Faix 1989) [2]. Oral nystatin is the most commonly used non-absorbable agent, followed by oral miconazole, which is also non-absorbable and an alternative to nystatin. Oral or intravenous fluconazole has been used both in the treatment of systemic infections [2], and more recently to reduce fungal colonization and infection [42].

For the Cochrane Neonatal Collaborative Review Group, Austin and Darlow [2] evaluated three trials, where oral antifungal agents were used as prophylaxis against systemic candida infection in preterm infants. They concluded that based on the available trials, the evidence does not yet support the oral administration of antifungal agents. In one of the analyzed trails, Sims et al. [65] reported a statistically significant reduction in the incidence of systemic fungal infection. Additionally, in none of these studies adverse effects were reported [2].

Conclusion

For the first time, a multimodal approach to NEC prophylaxis is reported. The present study used a combination of different strategies, all having shown to have some beneficial effect, but not having brought a clinical breakthrough in single administration studies. To prevent NEC effectively, the beneficial effects of human breast milk feeding, oral antibiotics, oral antifungal agents, and the administration of probiotics were combined together. In a homogenous group of preterm infants, using this protocol of multimodal NEC prophylaxis, the NEC incidence was as low as 0.7%, when prophylactic strategy was started within the first 24 h of life. This incidence was significantly lower when compared to a control group of infants without a primary prophylaxis strategy.