Introduction

Neonatal necrotizing enterocolitis (NEC) is a devastating disease that continues to remain poorly understood. NEC is an acquired gastrointestinal (GI) disease that mostly affects premature infants; the risk of development is the highest in infants born at <32 weeks’ gestation and with birth weight of <1500 g. The diagnosis was first described in 1960, and it has remained challenging to understand, prevent, and manage despite dedication to research. The prevalence of the disorder is about 7% among infants born weighing 500 to 1500 g. The rate of death resulting from complications of NEC is about 30%. In this discussion, we will review the basics of NEC, the role of the infant microbiome, and controversial risk factors, including antibiotic exposure, medications, anemia and blood transfusions, feeding regimens, and the use of probiotics.

NEC—Back to Basics

NEC, which is a common risk for premature infants, increases mortality and morbidity, including, but not limited to, severe neurodevelopmental outcomes and prolonged hospital stays. The development of NEC is multifactorial. It typically occurs between 2 and 6 weeks of life, with the highest risk at 29 to 33 weeks corrected gestation age. Classic signs and symptoms are nonspecific and include abdominal distension, poor tolerance of feeds, and hematochezia, along with subtle signs, such as activity change or increased occurrence of apnea. Early findings may be represented radiographically by notably dilated loops of bowel, paucity of bowel gas, fixed loops of bowel, and/or extraluminal air. However, an infant’s condition can deteriorate rapidly to a septic shock–like appearance, including hemodynamic instability and acute respiratory failure. The pathognomonic finding on abdominal radiography is pneumatosis intestinalis, or air tracking within the bowel wall. Management is only supportive, not specific, and includes bowel rest with intravenous (IV) nutrition, decompression of the abdomen, broad-spectrum antibiotics, and/or surgery. Surgery may involve placement of a peritoneal drain or laparotomy, with resection of the affected part of bowel. Treatment options depend on the severity of the disease in each neonate. Bell et al. described a staging system in the 1970s, and that system has been modified over the years, with improvements in neonatal management and understanding of other neonatal morbidities that may have caused confusion in the stratification of NEC earlier. The criteria are categorized into three stages and substages. Stage I is classified as suspected NEC—manifesting as temperature instability, apnea, and slight changes in feeding tolerance. Stage II is classified as definite NEC and includes stage I symptoms plus significant changes in abdominal examination with radiographic findings of pneumatosis or portal venous air. Stage III is classified as advanced NEC—manifesting as significant septic shock–like appearance, severe abdominal extension, and ascites or free air as seen on the radiograph ( Table 6.1 ).

Despite much research and further advancements in the neonatal intensive care unit (NICU) setting, management has remained essentially unchanged because of limited knowledge about the pathophysiology of NEC. We do not know how to specifically prevent or treat this disease. There is only a general understanding of NEC, and merely overall systemic supportive measures are provided to the sick neonate. The development of NEC is thought to be multifactorial in nature, and these factors include prematurity, changes in the neonatal microbiome, enteral feeding, early antibiotic exposure, lack of bowel perfusion, and/or medication exposure. A greater understanding of the pathogenesis of this disease is needed to tie together the seemingly disparate risk factors, to reliably predict patients at highest risk, and to provide specific therapies to prevent and treat this disease. A possible link is the microbiome of the newborn intestine. In this chapter, we will discuss a few of the most common theories about the triggers or influences that may increase the risk for NEC and their connection to the microbiome ( Fig. 6.1 ).

Neonatal necrotizing enterocolitis—clinical controversies and factors impacting the developing microbiome.

Microbiome

It was once thought that NEC could be caused by a direct insult to the intestine, whether from altered intestinal perfusion, infection, or other etiology. Research has revealed more areas of consideration and understanding of the dynamic community of the GI tract. The intestinal microbiome comprises all of the bacteria within the intestine, ideally with a balance among organisms, living symbiotically in a healthy environment.

Although bacteria appear to play a role, NEC is not an infection in the classic sense and cannot be isolated to one infecting agent or pathogen. In fact, in most cases, blood culture results are negative for infectious agents. A large cohort study in the 1980s reviewed all blood culture samples drawn from infants with NEC and noted that only about 30% of these culture samples showed bacterial growth. A more recent study in 2012 further demonstrated that <15% of blood cultures from infants with NEC were positive for bacteria.

The concept of one pathogen causing the insult has given way to questioning whether a community of organisms cause NEC. With advancements in technology during the 1990s, the microbiome began to be studied in depth. Molecular profiling of the microbiota, through sequencing of the highly conserved 16S small subunit bacterial ribosomal RNA genes, now allows identification of previously undetectable microbes. The composition of these microbes within the GI tract is dynamic, influenced by the environment, and undergoes changes in infancy. It was once thought that infants were born from a sterile environment because investigators were unable to grow bacteria from the amniotic fluid or from the surface culture samples from newborn infants directly after delivery. It was believed that bacteria begin to populate in the infant only after delivery. With further development of technologies, as previously discussed, the presence of bacterial species from the phyla Firmicutes, Tenericutes, Proteobacteria, Bacteriodes, and Fusobacteria have been found on the basal plate of the placenta (shared maternal–fetal interface). In multiple studies, bacteria have also been identified in low amounts within the umbilical cord and meconium of infants before initiation of feedings. It is, therefore, likely that the gut begins to colonize with bacteria in utero . Over time, the community becomes more complex so that by adulthood, the human intestinal microbiome consists of 10 to 100 trillion different microbial species.

A healthy microbiome is characterized by a high diversity of bacteria in balance, providing redundancy in function and protection against disease. Bacteria that are out of balance, a state termed dysbiosis , lose the ability to function at full capacity and to provide protection against disease. It is not surprising to find that the microbiomes of preterm and term infants differ. Several factors, including oxygen exposure altering the viability of anaerobic bacteria, the mode of delivery (vaginal delivery results in exposure to maternal vaginal flora versus cesarean section results in exposure to maternal skin flora), type of feedings, and exposure to medications, affect the development of the microbiome. For example, the microbiome of infants delivered via cesarean section does not harbor strict anaerobes and a large amount of facultative anaerobes, such as Clostridium species. The developing microbiome of all infants is influenced by all of these factors; in premature infants, however, development is being completed in the context of this microbiome. They are additionally exposed to unnatural settings, such as sterile hospital environments, medical instrumentation (nasogastric tubes, endotracheal tubes, intravenous lines, etc.), and feedings initiated at an earlier developmental stage than in term infants. All of these exposures lead to atypical development of the microbiome, which may place premature infants at a higher risk for developing NEC. Claud et al. compared the intestinal microbiome of healthy premature infants without NEC to that of healthy, vaginally delivered, breastfed, full-term infants. Stool samples were collected weekly over the first 8 weeks of life and sequenced by using the 16S rRNA process to document the evolving bacterial microbiome. Samples from full-term infants were found to be clustered with similar bacteria at all time points. In contrast, healthy premature infants revealed clustering at particular intervals including <2 weeks of age, 3 to 5 weeks of age, and >6 weeks of age. This suggests that the time course of the development of the microbiome may be important in premature infants.

Microbiota in healthy term infants is established in a stepwise process, starting with facultative anaerobes. Within a few weeks of life, Bifidobacterium species become highly prevalent. Diversity of the infant’s gut increases with time and undergoes a major shift during weaning. In contrast, preterm infants are exposed to prenatal and postnatal environmental insults while the microbiome is still undergoing development. The initial microbiome has low diversity. The microbiome is then influenced by the NICU environment, and community shifts may have an influence on the risk of NEC. In particular, a further decrease in species diversity and a bloom of gamma-proteobacteria has been demonstrated weeks before the actual clinical development of the disease.

In 2016 a case study published by Hourigan et al. highlighted the importance of the microbiome in twins, with only one of them developing NEC. Diamniotic and dichorionic twins (boy and girl) exhibited similar early NICU courses undergoing a course of antibiotics, total parental nutrition, and the initiation of feeds. Twin A had an uncomplicated case and was discharged from the NICU around the time of the expected due date; twin B had a complicated course, including medical closure of her patent ductus arteriosus and feeding intolerance, and ultimately developed NEC in the ileum with multiple perforations. Stools for each infant had been collected before twin B developed NEC, and each stool specimen was sequenced by using 16S rRNA sequencing technologies. Twin B had a much more diverse microbiome in earlier stages of life, but this diversity was abruptly lost weeks before the development of clinical NEC. Twin A had a prominence of Firmicutes in all samples (>95%), whereas twin B had both Proteobacteria and Firmicutes , with an increasing amount of Proteobacteria weeks before developing NEC. This case report highlighted the potential disparity between twins in microbiome development and clinical course despite their similar intrauterine and early environmental exposures. The ever-changing composition of specific bacteria may create an unbalanced microbiome with both increase in pathogens and decrease in protective species, leading to susceptibility to NEC. Can monitoring of the development of the microbiome in premature infants help identify those at higher risk of developing NEC? Could this be the strategy used in the near future to map those at a higher risk of developing NEC based on changes in the microbiome?

Furthermore, the microbiome plays a major role in immune system development and in protection against disease. The microbiota promotes epithelial cell regeneration, modulating epithelial permeability, as well as promoting angiogenesis, remodeling the intestinal vascular system. Preterm infant dysbiosis may contribute to gut dysmotility, immature barrier function, and poorly controlled immune responses, thus increasing the risk of NEC development.

As an example, growth is a parameter that can be clinically monitored in premature infants as a sign of health. In 2015 a study used a humanized gnotobiotic mouse model to investigate the functioning of the microbial community in preterm infants. Microbiota samples extracted from stools from a premature infant with appropriate growth and an infant with poor growth, both with the same gestational age and at <2 weeks of life, were gavaged to gnotobiotic pregnant dams. Both microbiomes were quite similar at the phylum level. The delivered pups acquired the designated microbiota from each infant naturally through birth and nursing from the newly colonized dams. The pup with the microbiota from the infant with poor growth was found to have significant upregulation of genes related to inflammatory response and the innate immunity. There was an increase in the baseline serum inflammatory cytokine levels even without an actual inflammatory insult. The pup with the microbiota from the properly growing infant showed downregulation of these genes. This finding suggests that a particular microbiome does modulate inflammatory responses, potentially protecting the premature infant and supporting healthier development. Furthermore, the pups from dams receiving the stools from the normally growing preterm infant had statistically significant longer small intestines, along with greater villus height and crypt depths. There were also more proliferating cells and decreased apoptosis along all sections of the intestinal tract in comparison with those pups from dams with microbiota from poorly growing preterm infants. The differentiation of cells differed significantly, including higher concentrations of goblet cells and Paneth cells within the microbiota of appropriately growing preterm infants. All these improved variables indicated an overall process of protection within the intestinal tract and better-developed mucosal defense and function. A healthy microbiome enhances intestinal development, protects against disease, and is influenced by the environment. What routine aspects of neonatal care influence the microbiome and NEC risk?

Antibiotics

Medications may disrupt microbiome balance. Antibiotics are the most common medications used in the NICU. Antibiotics have been investigated as a preventative measure against NEC. In the 1990s, double-blind randomized control trials, such as that by Siu et al., studied the use of oral vancomycin versus placebo in the prevention of NEC. Of the 71 premature infants included in the study by Siu et al., 13% developed NEC in the vancomycin group compared with 28% of those who received the placebo. Despite these significant results, prophylactic antibiotics for NEC have not been widely studied because of concerns about development of resistant bacteria. Furthermore, despite their possible effectiveness, antibiotics can have a significant impact on the microbiome. Most premature infants receive broad-spectrum antibiotics covering the most common neonatal pathogenic bacteria to prevent infection. However, studies have demonstrated alterations in microbiome balance in premature infants exposed to antibiotics. La Rosa et al. analyzed fecal samples from 58 infants over the course of their stay in the NICU and identified overall a common initial presence of Bacillus species transforming to higher amounts of Clostridia species weeks later. During this time, abrupt changes in bacterial species were noted in the microbiome. It is noteworthy that antibiotics were given to these neonates around the time of these identified abrupt changes.

Prolonged use of therapeutic agents influence initial intestinal colonization, and thus the length of antibiotic therapy may increase morbidity. A study by Cotten et al. reviewed 5693 extremely low–birth weight infants admitted to 19 different centers; 96% of infants were treated with a combination of two antibiotics with a median of 3 to 9.5 days of therapy, with more than 50% receiving more than 5 days of treatment. Longer durations of initial empirical antibiotic treatment were more likely to be associated with NEC and/or death compared with shorter durations. There was a 4% increase in the risk of NEC or death with each additional day of empirical antibiotic coverage. The increase in the risk of NEC alone almost doubled to a 7% increase for each additional day of coverage. The risk of NEC and/or death and the risk of death were increased with initial empirical antibiotic treatment lasting >5 days. Research on the specific effects of antibiotic use on microbiome development in preterm infants is ongoing, and more focus is needed before making changes to practice. However, available data indicate the need for caution regarding the potential adverse effects of certain empirical treatments.

Antireflux Medications

Another highly controversial topic is the use of antireflux medications, including proton pump inhibitors (PPIs) and histamine-2 (H2) blockers in the NICU. Although it is recommended that these medications be reserved for neonates who have evidence of pathologic exposure to acid reflux episodes, these medications tend to be overused. Two thirds of neonatologists use antireflux medications to treat apnea despite limited data demonstrating any association between reflux and cardiorespiratory events. Furthermore, these medications have potential adverse effects. Gastric acid inhibitors may increase the risk of NEC by altering the gastric pH, thus manipulating the protective properties of stomach acid and causing alteration in bacterial colonization. Physiologically, gastric pH is around 1.4 in adults. Any pH <4 has bactericidal effects on ingested acid-sensitive bacteria. Studies in rats and humans have shown a change in lower intestinal bacteria after exposure to a PPI increasing gastric pH to >4. Facultative and obligate anaerobic bacteria belonging to seven genera— Prevotella , Atopobium group, Clostridium coccoides group, Veillonella , Lactobacillus group, Enterobacteriaceae , and Bacteroides fragilis— were found in much higher levels in rats that received H2 blocker treatment compared with rats not exposed to a gastric acid inhibitor. The study also compared the use of a PPI versus an H2 blocker and found that use of a PPI had a statistically significant dose-dependent effect, increasing the populations of both Lactobacillus and Veillonella species. Interestingly, these two bacterial species originate in the oropharyngeal region, suggesting that failure of the gastric acid barrier possibly increased the load of bacteria entering the intestines.

These medications also delay gastric emptying and thin the viscosity of mucus, allowing easy bacterial overgrowth. Without a strong acidic environment, the immature humoral immune system cannot function at full capacity. The process of chemotactic response and leukocyte reactions to antigens is limited by the lack of the necessary acidic environment. There is decreased adhesion of leukocytes to endothelial cells in environments with a higher pH. There is reduction in bactericidal killing of microbes, and phagosomes no longer function, and this allows bacterial overgrowth. Wandall investigated the effects of omeprazole on polymorphonuclear neutrophil chemotaxis, superoxide generation, and degranulation and translocation of cytochrome b, which are important immunologic defenses. This study demonstrated that a higher concentration of omeperazole leading to a more alkalotic environment inhibited this process. With changes to pH, there is a change in bacterial species found within the microbiome, as well as inhibition of natural immune defenses, possibly leaving the intestinal mucosa vulnerable to pathogenic changes.

In August 2012, More et al. evaluated the association between exposure to H2 inhibitors and the incidence of NEC and infections in preterm infants. Guillet et al. evaluated 10,903 infants, of whom 72.2% were on H2 blockers. In the study by Terrin, of the 274 preterm infants included, 33.2% were on H2 blockers. The overall incidence of NEC was 7.1% and the risk of NEC was 6.6-fold higher in infants on H2 blockers. Overall, H2 blockers affect the protective barrier of the gastric lining and allow for bacterial overgrowth, altering the microbiome and increasing inflammation within the mucosal lining. These alterations may increase the risk of NEC.