Part 1. Current Controversies in the Understanding of Necrotizing Enterocolitis

Understanding the Pathogenesis

The precise pathogenesis of NEC is elusive. For a brief review of GI anatomy and physiology, see Sidebar 2. As early as 1975, the triad of GI mucosal injury, presence of bacteria in the gut, and introduction of a metabolic substrate (eg, feedings) was proposed as the etiology of bowel necrosis^[3] (Fig 2). Ongoing investigation of the role of ischemic mucosal injury in NEC has now progressed to the molecular level.

Histologic findings in NEC include GI mucosal ulceration and edema, hemorrhage, vascular congestion, and in advanced cases, full-thickness coagulation necrosis with perforation; all of these findings are consistent with ischemia.^[1,4,8,21] Early mucosal lesions lead to marked broadening of the mucosal villi, with advanced NEC causing bleeding and sloughing of surface epithelium. Hydrogen gas pockets, produced by gas-forming organisms, arise in the submucosa and subserosa. Severe necrosis may completely obliterate the villi (Fig 3A and B).

Maternal cocaine use during pregnancy may result in vasoconstriction and GI ischemia from cocaine's -adrenergic action, predisposing infants to NEC.^[22] An increased incidence of surgical NEC, mortality, and massive GI gangrene has been found in cocaine-exposed infants (Fig 4). There were no differences in rates of perinatal asphyxia, hypotension, or UAC placement.

Conflicting evidence exists regarding the role of maternal preeclampsia and NEC. Because preeclampsia exposes the fetus to a chronic reduction in uteroplacental blood flow, the potential for decreased GI perfusion exists. A recent study (n = 242) measured umbilical artery end-diastolic flow in 41 neonates who developed NEC.^[23] Contrary to historical data, no association between absent end-diastolic flow with or without abnormal fetal heart rate patterns at birth and the incidence of definitive NEC was found.^[24,25]

During a sustained underwater dive with suboptimal cardiac output, certain mammals demonstrate an autonomically mediated diving reflex, resulting in redistribution of blood flow away from the GI tract and toward vital organs. In 1969, Lloyd proposed a similar phenomena in asphyxiated neonates as a major factor in the development of NEC.^[26] Correspondingly, the most common sites for NEC, the distal ileum and proximal colon, lie between watershed areas of the superior and inferior mesenteric arteries.^[21] These watershed areas represent potential zones of decreased perfusion during ischemia.

Recent information weakens the connection between bowel ischemia, NEC, and the diving reflex postulate. Although the relationship seems rational, the underlying premise is false. Lloyd's theory was that decreased GI blood flow impaired local tissue oxygenation, leading to bowel necrosis. In most cases, however, decreased GI blood flow leads to microvasculature compensation, preserving tissue oxygenation.^[21] Only severe ischemia disrupts the process.

Current evidence suggests that although asphyxic events impair GI perfusion, the damage is transient and not great enough to cause GI necrosis without some additional factors.^[21] Hypoxic and ischemic damage to the bowel may actually be a secondary event aggravated by components such as inflammatory mediators, immature GI vascular control, and chemical irritation.^[8]

Epidemiologic case control studies suggest that hypoxic-ischemic insults are not primary risk factors in the development of NEC. Prematurity is the only independent risk factor consistently identified.^[2,3,8] Additionally, many cases of NEC occur in convalescing preterm neonates with no known antecedent hypoxic-ischemic stressors (eg, low Apgar scores, mechanical ventilation, umbilical catheters, significant apnea).^[3,8] Further research is needed to answer questions regarding the association of ischemia and NEC.

The impact of inflammatory mediators on a variety of disease processes is under intense investigation because of their influence on regional blood supply, capillary permeability, and leukocyte migration.^[27] Inflammatory mediators are likely to be a component of GI mucosal injury's role in the development of NEC.^[4,27] Inflammatory cytokines, proteins produced by macrophages in response to infection, worsen any hypoxic cellular damage.

The cytokines interleukin-6 (IL-6), produced in response to circulating bacteria, and tumor necrosis factor (TNF), a key mediator in septic shock, have been examined in neonates with NEC.^[28] Although TNF did not differ between groups, IL-6 levels were elevated in neonates with bacterial sepsis and NEC compared with those with isolated bacterial sepsis or those with NEC without bacteremia.^[28] The majority of infants with NEC have negative blood cultures; thus, testing for IL-6 elevations, which measures a systemic response to severe bowel injury, may not be clinically useful.^[29]

Elevations in the proinflammatory cytokines IL-8 and IL-10 have correlated significantly with both NEC severity and time after onset of symptoms.^[29] This test may have the potential to help differentiate between infants likely to recover from NEC with little intervention from those at risk for severe disease.^[29]

Another inflammatory mediator implicated in the pathogenesis of NEC is platelet activating factor (PAF), an endogenous phospholipid messenger. PAF levels were measured in 164 at-risk neonates with a mean gestation of 30 weeks before onset and during progression of NEC.^[30] Higher PAF levels were found in 11 infants who developed NEC.

Immaturity of GI mucosal defense mechanisms against invading organisms may play a role in the pathogenesis of NEC ( Table 3 ). Lysozyme (muramidase), a bactericidal enzyme, appears early in fetal life and is secreted in the bowel by Paneth cells in the small intestine. Delay in Paneth cell maturation may impair production of lysozyme and defensins, both potent antimicrobial factors.

Lysozyme deficiency was evaluated by comparing infants with small bowel mucosa (n = 10 infants with NEC; mean gestational age, 31.3 weeks) with control infants.^[31] All control samples stained positive for lysozyme. However, no lysozyme was found in 9 of 10 NEC samples, strongly indicating an absence of lysozyme in the GI mucosa of preterm and term neonates with NEC. A delay in GI maturation is suggested as a contributing factor in NEC; the potential to use Paneth cell inducers as prophylaxis against NEC has been proposed.^[31]

Another component believed to contribute to GI immaturity is enteric human defensins. Defensins are a family of antibiotic peptides with a broad range of antibiotic activity, found in circulating neutrophils and epithelial cells of the bowel. Low levels of human defensin 5 (HD5) are present in the neonatal bowel at 24 weeks of gestation; amounts increase as gestation continues.^[32] Elevated levels of messenger RNA HD5 in NEC GI mucosal tissue samples were found in 6 infants with NEC, compared with controls.^[32] The process of NEC is proposed as a trigger for messenger RNA HD5 production, indicating a role for this GI defensin in the pathophysiology of the NEC.^[32]

Epidermal growth factor (EGF) appears to be important during GI development and mucosal repair.^[33] Salivary glands are a major source of EGF production. Preterm neonates have low levels of EGF, which increase with gestational age. The use of exogenous EGF administration to salvage neonates who are terminally ill with NEC has been suggested.^[33]

A recent analysis compared EGF levels in saliva, serum, and urine in premature infants with and without NEC.^[33] EGF levels were significantly lower in saliva and serum of the NEC group, who were also more preterm, supporting EGF as a possible factor in the pathogenesis of NEC. Future development of clinically useful markers of EGF levels may guide practice. The potential for exogenous EGF administration as an NEC prevention measure calls for further exploration and testing.

In the GI tract, the T lymphocyte C-KIT plays an important role in local immunity; C-KIT numbers increase with inflammation. C-KIT cells monitor alterations in cell membranes infected by bacteria or virus, and destroy the invading organism. Because invading organisms or normal GI bacterial flora extend into the bowel lumen in 40% of NEC cases, the role of C-KIT warrants exploration.^[34] Postmortem specimens from 9 neonates with proven NEC were compared with agematched control specimens.^[34] Controls displayed normal distribution of C-KIT cells in the GI submucosa. Infants with NEC had significantly fewer C-KIT cells, despite the presence of local inflammation. Lower CKIT numbers may impair local gut immunity, allowing for bacteria, even normal flora, to overgrow, initiating the inflammatory cascade leading to NEC.

NEC rarely occurs in utero, suggesting that bacterial colonization is associated with the disease.^[8,35] The absence of NEC in the sterile in utero bowel and in stillborn infants strengthens the role of infection in its pathogenesis.^[35,36] Others assert that by definition (via NEC staging criteria), bacteria are a necessary ingredient in the disease process.^[37] The radiographic hallmark of NEC, pneumatosis intestinalis, signifies intramural gas, which is believed to be a result of hydrogen gas produced during the bacterial fermentation of enteral feedings^[36] (Fig 5). Although rare, infants who were never fed and develop NEC seldom have pneumatosis.^[37]

Breath hydrogen levels increase before the clinical onset of NEC.^[36] Increased urinary D-lactate excretion in neonates with NEC is probably a result of increased bacterial activity.^[36] Many believe that the systemic illness associated with NEC results from bacterial toxins or from strains of organisms that are high carbohydrate fermenters, leading to the lowering of bowel pH with resultant injury.^[38]

Most NEC occurs sporadically; however, case clustering has been reported.^[1] Clustered cases tend to affect larger infants and those with better Apgar scores and fewer perinatal complications, and are associated with decreased case-mortality rates.^[2] Case clusters are often associated with nonspecific GI disturbances among other neonates and the NICU staff, leading many to speculate that these outbreaks are infectious in origin.^[4]

Although NEC clusters or epidemics may provide persuasive evidence for the role of infection in pathogenesis, a unifying pathogen remains elusive.^[37,38,39] A variety of organisms, including Enterobacteriaceae, Clostridia, coagulase-negative staphylococci, Escherichia coli, and Klebsiella, have been associated with NEC. Even during outbreaks, typically no single pathogen is implicated.

Case reports of NEC caused by nosocomially acquired echovirus have been published.^[40] Overcrowding is associated with outbreaks, supporting the hypothesis of an infectious agent with a higher attack rate during crowding.^[2] It also reflects an increase in NICU census, resulting in a greater number of at-risk neonates.

Despite the failure to establish a direct relationship between a specific microbe and NEC, the presence of bacteria appears crucial for the development of NEC.^[36,37,38] Many organisms isolated from stool, blood, or peritoneal fluid in neonates with NEC are also commonly found in the bowel. Exposure to broad-spectrum antibiotics shortly after birth may markedly alter the GI flora. This could result in an intestinal environment with large numbers of bacteria more conducive to the development of NEC.

The microflora of the duodenum of VLBW infants is initially sterile or contains gram-positive organisms. Colonization changes with time^[36]; gram-negative organisms, primarily E. coli, Klebsiella, and Enterobacter species, appear once enteral feedings are started, and predominate after 15 days of age.^[36] The high incidence of Enterobacteriaceae may be related to decreased secretion of gastric acid, the normal line of defense against small bowel colonization. Limited GI motility and bowel stasis common in premature infants may also contribute. Thus, functional GI immaturity in the presence of gram-negative organisms may trigger the inflammatory process, resulting in tissue injury and ultimately NEC.^[41]

Translocation is the ability of bacteria to cross bowel mucosal epithelium and enter the lymphatic system or bloodstream. Strains of E. coli not previously recognized as pathogenic are able to translocate across mucosal cell layers.^[41] The exact mechanisms responsible are ambiguous, but likely do not involve initial tissue injury. The ability of bacteria to cross epithelial cell layer is a crucial first step in the cascade of events leading to NEC.^[41] It is unclear whether bacteria and substrate interact in the immature GI tract to damage the bowel or if stressors damage the immature GI tract to produce an environment in which bacteria proliferate in the presence of available substrate.

Adv Neonatal Care. 2003;3(3) © 2003 W.B. Saunders

Cite this: Part 1. Current Controversies in the Understanding of Necrotizing Enterocolitis - Medscape - Jun 01, 2003.