113: Necrotizing Enterocolitis

Necrotizing enterocolitis (NEC) is an ischemic and inflammatory necrosis of the bowel primarily affecting premature neonates after the initiation of enteral feeding.

NEC is predominantly a disorder of preterm infants, with an incidence of 6–10% in infants weighing <1500 g, with the highest incidence in the most premature infants. NEC can also occur in term infants, many of whom have preexisting medical conditions.

Multifactorial theory has been suggested in which several risk factors, including prematurity, formula feeds, ischemia, and bacterial colonization, interact to initiate mucosal damage via a final common pathway involving activation of an inflammatory cascade. Mucosal damage results in invasion of the bowel walls by gas-producing bacteria, resulting in intramural gas accumulation (pneumatosis intestinalis). This sequence of events may then progress to transmural necrosis or gangrene of the bowel wall and finally to perforation and peritonitis.

1. Prematurity. There is an inverse relationship between gestational age and risk for developing NEC. While most preterm infants develop NEC at postmenstrual age (PMA) of 30–32 weeks, various factors resulting from premature birth places them at increased risk for NEC. These may involve immature mucosal barrier, mucosal enzymes, and various gastrointestinal (GI) hormones. Premature infants may have an imbalance between pro- and anti-inflammatory factors and thus have increased activation of inflammatory mediators and decreased inactivation of specific mediators like platelet activating factor (PAF), which has been linked to NEC. Abnormal toll-like receptor 4 (TLR 4) signaling in the premature intestine and increased activation of nuclear factor-κB may play a role in the pathogenesis of NEC. An inability to effectively regulate the intestinal microcirculation and differences in bacterial colonization may also make preterm infants more susceptible to NEC.

2. Microbial colonization. NEC has not been shown to occur in germ-free animals. While bacterial and viral pathogens including Escherichia coli, Klebsiella spp, Clostridium spp, Staphylococcus epidermidis, rotavirus, and enterovirus have been implicated, no single organism has been consistently associated with NEC. Blood cultures are positive in only 20–30% of cases. While colonization by normal gut flora supports the intestinal mucosa through toll-like receptors, pathological bacteria induce inflammation and apoptosis by signaling pathways such as nuclear factor-κB. The growth of these noncommensal bacteria may also result in endotoxin release, leading to mucosal damage.

3. Enteral feedings. NEC is rare in unfed infants, and 90–95% infants with NEC have received at least 1 enteral feed. Enteral feeding provides necessary substrate for proliferation of enteric pathogens. Hyperosmolar formula/medications may alter mucosal permeability and cause mucosal damage. Short-chain fatty acids produced as a result of colonic fermentation (due to deficient lactase activity in premature infants) may add to the damage.

   Breast milk significantly lowers risk of NEC. It has the benefit of providing immunoprotective as well as local growth-promoting factors not available with commercially prepared formulas. Early initiation of small-volume feeds with human milk for a few days followed by slow advancement of feeds (about 20 cc/kg/d) has been shown to decrease the incidence of NEC. This can be effectively accomplished by using weight-based feeding guidelines.

4. Circulatory instability. During periods of circulatory stress, blood is diverted away from the splanchnic circulation (diving reflex). The resulting intestinal ischemia followed by reperfusion may lead to bowel damage. Imbalance between vascular dilator and constrictor molecules leading to defective splanchnic blood flow autoregulation in newborns may contribute to the injury. Infants with NEC have been shown by Doppler flow to have higher flow resistance in the superior mesenteric artery on the first day of life. A diminished blood supply to the gut in infants exposed to maternal cocaine may also increase the risk for NEC.

5. Maternal cigarette smoking. A recent study has found an association between maternal cigarette smoking and development of NEC in the infant. The underlying mechanism may be the effect of maternal smoking or nicotine on blood vessel development in the fetal gastrointestinal tract.

6. Congenital heart disease. Infants with hypoplastic left heart or univentricular heart disease with or without arch obstruction, truncus arteriosus or aortopulmonary window, congestive heart failure and infants who have had systemic-to-pulmonary shunts placed are at increased risk for NEC. Infants with symptomatic patent ductus arteriosus are also at higher risk for NEC. A common feature of many of these conditions is decreased mesenteric perfusion from diastolic steal. Hypoxemia from cardiac failure or cardiac procedures may add to this risk.

7. Polycythemia and hyperviscosity syndromes. These increase the risk for NEC by diminished perfusion and intestinal ischemia in watershed areas of the GI tract.

8. Blood transfusion. A significant association between red blood cell transfusion and NEC has been reported in recent retrospective studies with about 25–35% cases of NEC occurring within 48 hours of packed red blood cell (PRBC) transfusions. The pathogenic mechanism is uncertain but the following mechanisms have been proposed:

   1. Immunologic mechanism. Transfusion-related gut injury (TRAGI) may be similar to the transfusion-related lung injury (TRALI) described in adults with immunologically mediated damage to the gut. Impaired maturation of T cells or priming of neutrophils from prior transfusions may play a role.

   2. Anemia. Transfusions are given for symptomatic anemia. Anemia may impair the oxygen-carrying capacity of blood, leading to reduced oxygen delivery to the intestine resulting in injury. In 1 retrospective case-control study, NEC cases had lower hematocrit (Hct), and each one-point decrease in nadir Hct was associated with a 10% increase in the odds of NEC. Thus, withholding treatment of severe anemia may increase the risk for NEC. The critical Hct level where the risk of NEC from anemia outweighs the risk of blood transfusion has yet to be determined.

   3. Storage effects. Stored erythrocytes have depleted nitric oxide levels and could act as nitric oxide sink in the microcirculation predisposing to vasoconstriction and ischemic insult. The 3 mechanisms described previously may not be mutually exclusive and may all play a role in the pathogenesis of NEC. Feeding prior to, during, and after PRBC transfusion remains controversial. One institutional experience with withholding feeds prior and during PRBC transfusion showed a significant reduction in the incidence of NEC in their neonatal intensive care unit (NICU). Although many centers have instituted different feed-holding policies with blood transfusions, more prospective research is needed before uniform recommendations can be made.

While term infants who develop NEC often have underlying illnesses predisposing to NEC and are often diagnosed in the first week of life, most premature infants who develop NEC are between 14 and 20 days of age or 30–32 weeks' postmenstrual age. The early clinical presentation may include feeding intolerance, increased gastric residuals, and blood in stools. Specific abdominal signs include abdominal distension, tenderness, abdominal skin discoloration, emesis, and bilious drainage from nasogastric tube. Nonspecific signs include symptoms and signs of neonatal sepsis including increased apnea/bradycardia episodes, temperature instability, hypotension, and circulatory shock.

The clinical course of NEC is variable. While about 30% may have a mild presentation that responds to medical treatment, about 7% may have a fulminant course with rapid progression to NEC totalis, septic shock, severe metabolic acidosis, and death. The modified Bell's staging criteria is often used to classify NEC according to clinical and radiographic presentations.

1. Stage I: suspected NEC

   1. Systemic signs. Nonspecific, including apnea, bradycardia, lethargy, and temperature instability.

   2. Intestinal findings. Feeding intolerance, recurrent gastric residuals, and abdominal distension.

   3. Radiographic findings. Normal or nonspecific.

2. Stage II: proven NEC

   1. Systemic signs. Include Stage I signs plus abdominal tenderness and thrombocytopenia.

   2. Intestinal findings. Prominent abdominal distension, tenderness, bowel wall edema, absent bowel sounds, and gross bloody stools.

   3. Radiographic findings. Pneumatosis with or without portal venous gas.

3. Stage III: advanced NEC

   1. Systemic signs. Respiratory and metabolic acidosis, respiratory failure, hypotension, decreased urine output, shock, neutropenia, and disseminated intravascular coagulation (DIC).

   2. Intestinal findings. Tense, discolored abdomen with spreading abdominal wall edema, induration, and discoloration.

   3. Radiographic findings. Pneumoperitoneum (see Figure 11–22).

A high index of suspicion must be maintained for any infant with a combination of risk factors enumerated under Section IV.

1. Clinical diagnosis. NEC is a tentative diagnosis in any infant presenting with the triad of feeding intolerance, abdominal distension, and grossly bloody stools. Alternatively, the earliest signs may be identical to those of neonatal sepsis.

2. Laboratory studies. The following studies should be performed and repeated as necessary:

   1. Complete blood count (CBC) with differential and platelets. The white blood cell (WBC) count may be normal but is frequently either elevated with increased left shift or low (leucopenia). Thrombocytopenia is often seen.

   2. C-reactive protein (CRP) correlates with the inflammatory response. Since initial CRP may be normal, serial CRP levels done at 12- to 24-hour intervals are more useful.

   3. Blood culture for aerobes, anaerobes, and fungi (Candida spp).

   4. Stool cultures for rotavirus and enteroviruses.

   5. Electrolyte panel. Electrolyte imbalance, including hypo- and hypernatremia, as well as hyperkalemia, are common.

   6. Arterial blood gas measurements. Metabolic or combined acidosis may be seen.

   7. Coagulation studies include prothrombin time (PT), partial thromboplastin time (PTT), fibrinogen, and fibrin degradation products (d-dimer). A rise in PT, PTT, and fibrinogen degradation products indicates DIC, a frequent finding in infants with severe NEC.

3. Imaging and other studies

   1. Flat plate radiograph of the abdomen

      1. Supportive for NEC. Abnormal bowel gas pattern, ileus, fixed sentinel loop of bowel, or areas suspicious for pneumatosis intestinalis.

      2. Confirmatory for NEC. Pneumatosis intestinalis and intrahepatic portal venous gas (in the absence of umbilical venous catheter). (See Figures 11–23 and 11–24.)

   2. Lateral decubitus and cross-table lateral studies of the abdomen. Presence of free air is indicative of intestinal perforation. Serial radiographic studies of the abdomen should be obtained every 6–8 hours in the presence of pneumatosis intestinalis or portal venous gas to look for pneumoperitoneum as they are at risk for perforation within 48–72 hours. Serial radiographs may be discontinued with clinical improvement after 48–72 hours.

   3. Abdominal sonography. May be useful in the presence of nonspecific clinical and radiologic findings or in infants with NEC not responding to medical management. Ultrasound can detect intermittent gas bubbles in the liver parenchyma and the portal venous system that are not detected on abdominal radiograph. Free gas and focal fluid collections can also be viewed by sonography. Color Doppler ultrasound (US) is useful in detecting bowel necrosis and perfusion.

   4. Mesenteric oxygen saturation. Recent studies have shown the possibility of using near-infrared spectroscopy (NIRS) to detect mesenteric oxygen saturations. This provides hope of early detection and real-time noninvasive monitoring for decreased bowel perfusion in infants who have NEC. This technique is still experimental.

The main principle of management for confirmed NEC is to treat it as an acute abdomen with impending or septic peritonitis. The goal is to prevent progression of disease, intestinal perforation, and shock. If NEC occurs in epidemic clusters, cases need to be isolated.

1. Medical management

   1. Nil per os (NPO) to allow gastrointestinal rest for 7–14 days (shorter course for Stage I NEC). Total parenteral nutrition (TPN) to provide basic nutritional needs.

   2. Gastric decompression with large-bore orogastric tube (Replogle) at low intermittent or continuous suctioning.

   3. Close monitoring of vital signs and abdominal circumference.

   4. Monitoring for gastrointestinal bleeding. Check all gastric aspirates and stools for blood.

   5. Respiratory support. Provide optimal respiratory support to maintain acceptable blood gas parameters. Progressive abdominal distension causing loss of lung volume may increase need for positive-pressure ventilation.

   6. Circulatory support. There may be excessive third spacing of fluid, which requires effective volume replacement. Inotropic support may be needed to maintain normal blood pressure.

   7. Strict fluid intake and output monitoring. Try to maintain urine output of 1–3 cc/kg/h. Fluid replacement to correct third space losses. Removal of potassium from IV fluids in the presence of hyperkalemia or anuria.

   8. Laboratory monitoring. Check CBC and electrolyte panel every 12–24 hours until stable. Obtain blood and urine culture prior to starting antibiotics.

   9. Antibiotic therapy. Treat with parenteral antibiotics for 10–14 days. Antibiotic regimen should cover pathogens that can cause late-onset sepsis in premature infants. Add anaerobic coverage if bowel necrosis or perforation is suspected. Reasonable antibiotic regimens include

      1. Vancomycin, gentamicin, and clindamycin (or metronidazole).

      2. Vancomycin and piperacillin/tazobactam.

      3. Vancomycin, gentamicin, and piperacillin/tazobactam.

      4. Term infants may be treated with ampicillin, gentamicin, and clindamycin.

   10. Monitoring for DIC. Infants in Stage II and III may develop DIC and require fresh frozen plasma and cryoprecipitate. Packed red blood cell transfusions and platelet transfusions may also be needed.

   11. Imaging studies. Obtain an abdominal flat plate with lateral decubitus of cross-table lateral studies every 6–8 hours in the acute stage (usually the first 24–48 hours) to detect bowel perforation.

   12. Surgical consultation is needed for confirmed NEC Stage II and III, especially when the condition is rapidly progressing or there is evidence of GI perforation.

2. Surgical management. A pneumoperitoneum is an absolute indication for surgical intervention. Relative indications for surgery include portal venous gas, abdominal wall edema and cellulitis (indicating peritonitis), fixed dilated intestinal segment by x-ray (sentinel loop), tender abdominal mass, and clinical deterioration refractory to medical management.

   1. Exploratory laparotomy. This involves examining the bowel and resecting the necrotic segments. A portion of viable bowel is used to create an enterostomy and mucous fistula. Reanastomosis takes place after 8–12 weeks. If NEC only involves a short segment of bowel or limited resection, primary anastomosis is used by some surgeons, avoiding complications associated with ileostomy and need for second surgery for reanastomosis. In situations of widespread areas of intestinal necrosis, abdomen may be closed after placement of a drain and reexplored later. A poor prognosis is associated with severe short-bowel syndrome, and foregoing further treatment may be considered.

   2. Peritoneal drainage (PD) placement. A small transverse incision is made at McBurney's point. Abdominal layers are bluntly dissected and a Penrose drain is threaded into the abdomen and secured.

      Two multi-center trials have shown use of PD and laparotomy in infants with bowel perforation to have similar mortality, need for total parenteral nutrition, and length of hospital stay. There has been concern for the requirement of delayed laparotomy in infants managed with PD. This has varied from 38% in the study by Moss et al and 74% in that by Rees et al, without affecting survival. PD is a relatively simple procedure and can be done with local anesthesia at the bedside. Hence, it is often used as a temporizing procedure in critically sick infants. However, this too has been questioned by Rees et al, and they have shown lack of improvement in physiologic measurements following PD. Currently, the optimal surgical management for infants with bowel perforation remains controversial and needs further study.

1. Human milk has been shown to prevent NEC. While mother's own milk is ideal, a meta-analysis of 5 randomized clinical trials of donor human milk versus formula suggests that human milk was beneficial and formula-fed infants had a 2.5 times increased risk of NEC. Another study showed that the risk of NEC decreased by a factor of 0.8 for each 10% increase in the proportion of human milk intake. The rate of NEC was also shown to be lower in infants receiving human milk fortifier compared to those receiving bovine milk–based fortifier.

2. Use of feeding protocol with initial period of trophic feeds followed by gradual advancement of feeds has been shown to decrease the incidence of NEC. A cautious approach to feeds in high-risk infants with circulatory compromise, congenital heart disease, or those receiving a blood transfusion is recommended.

3. Probiotics. There has been recent work promoting use of probiotics to decrease the risk for NEC in preterm infants. Probiotics may prevent NEC by promoting colonization of the gut with beneficial organisms, preventing colonization by pathogens, improving the maturity and function of gut mucosal barrier and modulation of the immune system. A meta-analysis of 11 published trials on use of probiotics showed a 30% reduction in the incidence of NEC with no significant adverse effects. However, additional studies regarding the safety of specific probiotic preparations, dosage, duration, and practicality of administration in this fragile population are awaited before routine treatment with probiotics can be recommended. Probiotics have not been approved by the U.S. Food and Drug Administration.

4. Prebiotics, or nutrients that enhance the growth of beneficial microbes, have been proposed as a preventive strategy. These include oligosaccharides, inulin, galactose, fructose, lactose, and others. While prebiotics enhance the proliferation of endogenous flora, their efficacy in prevention of NEC is unclear.

5. Avoidance of prolonged empiric antibiotic use. These alter the gut flora, promoting growth of pathogens, and should be avoided in premature infants. This is supported by a retrospective study that showed that extremely low birthweight infants receiving initial antibiotic course of >5 days had an increased risk of NEC or death.

1. Recurrence of NEC may occur in about 5% of cases.

2. Colonic strictures may occur in 10–20% cases and present with recurrent abdominal distension and persisting feeding intolerance. Contrast radiographic studies are usually diagnostic.

3. Short bowel syndrome may develop in infants undergoing extensive resection of bowel. The traditional limits of intestinal length for successful survival are at least 20 cm of viable small bowel remaining with an intact ileocecal valve, or 40 cm viable remaining small bowel with loss of ileocecal valve. Intestinal transplantation, with variable outcome, remains an option for some of these infants.

4. Parenteral nutrition–associated liver disease occurs more frequently in infants with surgical therapy for NEC.

1. Infants with NEC have an overall mortality of 12.5%. NEC with perforation is associated with mortality of 20–40%.

2. Infants with surgical NEC are at risk for significant growth delay and adverse neurodevelopmental outcomes.