key: cord-022216-k4pi30sd
authors: Kliegman, Robert M.
title: Neonatal necrotizing enterocolitis
date: 2009-05-15
journal: Pediatric Gastrointestinal and Liver Disease
DOI: 10.1016/b978-0-7216-3924-6.50047-0
sha: 
doc_id: 22216
cord_uid: k4pi30sd

nan

Neonatal necrotizing enterocolitis (NEC) is a disease of unknown origin that predominantly affects premature infants in level II, or more often level III, neonatal intensive care units during the infant's convalescence from the common cardiopulmonary disorders associated with prematurity. [1] [2] [3] [4] [5] [6] NEC is the most common and most serious acquired gastrointestinal disorder among hospitalized preterm neonates and is associated with significant acute and chronic morbidity and mortality. 1, 7 Indeed, NEC is the most common cause of gastrointestinal perforation (followed by isolated idiopathic focal intestinal perforation) and acquired short bowel syndrome among patients in the neonatal intensive care unit. 1, 7, 8 It has been estimated that in the USA there are approximately 2000 to 4000 cases of NEC annually. 1, [9] [10] [11] NEC is one of the most common nosocomial neonatal diseases; additional hospital charges for NEC were estimated to be $6.5 million dollars at one neonatal system, with an average additional cost of $216 000 per survivor. 5

NEC is a disease that affects premature infants during their convalescence from other diseases of immaturity. 1, 9 It is unusual to see NEC during the acute phase of respiratory distress syndrome, hypoxic-ischemic encephalopathy (birth asphyxia), heart failure from a patent ductus arteriosus and other acute neonatal disease processes. 1, 12 Although NEC is noted predominantly in premature neonates, approximately 10% of cases occur in nearly full-term or full-term infants whose preceding risk factors have included polycythemia, cyanotic heart disease or heart disease producing low cardiac output (before or after surgery), chronic diarrhea, endocrine disorders (hypothyroidism, panhypopituitarism, congenital adrenal hyperplasia) or a prior anatomic obstructive gastrointestinal malformation (volvulus or gastroschisis). 1, 2, 13 NEC occurs predominantly in neonates after the onset of enteral alimentation, as 95% of affected infants have been fed for various amounts of time. 1, 12, 14 The condition usually develops during the first 2 weeks of life in relatively well neonates who have been fed enterally, but it may be delayed for 90 days in infants of very low birthweight given nothing by mouth for long periods. 1 NEC tends to develop later among infants whose birthweight was less than 1000 g. Thus, the more immature the infant at birth, the later the possible onset of NEC. Such findings suggest that immaturity of gastrointestinal function is a major risk factor. 1, 12, 14 Many case-control studies have suggested that immaturity is the only readily identifiable risk factor for the development of NEC. 1 Immature function of host defense, gastrointestinal motility, digestion, healing and mucosal integrity (permeability), and intestinal circulation may contribute to the pathogenesis of NEC (see below). 14, 15 The incidence of NEC varies among neonatal intensive care units, and some units have no cases. The overall incidence varies from 3% to 5% of all neonatal intensive care admissions. 1, 9 The incidence among very-low-birthweight (less than 1500 g) infants who survive long enough to be fed approaches 10-15%. The incidence is greatest among infants weighing between 500 and 750 g at birth (13-20%) ; it declines to approximately 1-3% among infants weighing more than 1750 g. 1, 4 The incidence is affected by periodic outbreaks or epidemics of NEC that become superimposed on an endemic case background. Because infants of 500-750 g at birth represent a very small proportion of infants in neonatal intensive care units (less than 2% of all births), the mean birthweight of infants affected with NEC is between 1350 and 1500 g and the mean gestational age is between 30 and 32 weeks.

There is usually no association between NEC and sex, race (other than the increased incidence of low birthweight among African Americans), or inborn vs transport status. 1 NEC has very rarely been reported in successive pregnancies and among siblings in multiple gestations. Because twins and triplets are often born prematurely, they seem to be over-represented among babies with NEC. 1 In contrast to what would be expected, the incidence of NEC among twins appears to be higher in the well, first-born twin rather than in the second twin, who usually has a lower Apgar score and a more complicated hospital course.

NEC is a disease of surviving neonates. Most affected patients are considered 'gainers and growers' who, prior to the onset of NEC, had few if any manifestations of gastrointestinal dysfunction, and few apparent sequelae of the previous diseases of prematurity from which they had been recovering. 12 In many but not all instances, the affected pre-term infants had respiratory distress syndrome days to weeks before the onset of NEC. With the advent of surfactant therapy for respiratory distress syndrome, more infants may survive this disorder and are now at risk for NEC during their convalescence. 12 Because NEC is seen in premature infants, many factors associated with prematurity were thought to be risk factors for NEC. These were related to diseases, procedures or complications that may produce gastrointestinal ischemia, infection or altered digestion. Mucosal injury following a combination of these risk factors was thought to produce NEC. Ischemic risk factors included birth asphyxia, respiratory distress syndrome, hypoxia, hypotension, patent ductus arteriosus, polycythemia, anemia, umbilical artery, catheter placement and exchange transfusion. Because these risk factors were identified as part of a profile among affected infants, they were thought to contribute to the pathogenesis of NEC. Although each may contribute to mucosal injury, case-control studies suggest that these processes are equally prevalent among affected patients and unaffected controls. 1 Therefore, these risk factors merely describe the low-birthweight population and are no longer thought to be direct contributing factors in the pathogenesis of NEC (see below). Epidemiologic investigations emphasize that the predominant risk factor is prematurity, with associated immature host defense and gastrointestinal function.

The pathologic appearance of NEC has demonstrated varying features of inflammation and coagulation necrosis. 16 The latter has been interpreted as being due to ischemia, but evidence also suggests that inflammatory mediators can produce lesions with coagulation necrosis similar to that following ischemia. 17, 18 Severe NEC is a transmural process, involving all four layers of the bowel wall. NEC often involves the terminal ileum, ileocecal area and ascending colon, and rarely only the rectal mucosa. Eosinophilia may be noted in the rectal mucosal biopsies of patients with mild disease; this may represent only the early initiation phase of the illness, which later becomes masked by complicating peritonitis, secondary bacterial invasion and neutrophilic infiltration. Eosinophilic intestinal infiltrates are not always suggestive of an allergic process and may be a non-specific response.

Superficial inspection of affected tissue reveals mucosal ulceration, hemorrhage, edema, and submucosal or subserosal gas-filled cysts typical of pneumatosis intestinalis ( Fig. 43.1 ). Approximately 50% of infants have involvement of both the small and large intestine (distal ileum and proximal colon), 25% have only colonic disease and 25% only ileal lesions. 1, 16 Involvement of the jejunum, stomach or entire length of bowel from the ligament of Treitz to the rectum is less common (seen in 10% of surgical cases); the latter is present in fatal cases of NEC (also known as pan-NEC or NEC totalis). Approximately equal numbers of infants have either continuous segments of involvement or skip lesions.

Histologic examination of resected tissue reveals coagulation necrosis in approximately 90%, inflammation in approximately 90% and the presence of eosinophils (15%), ulceration (75%), hemorrhage (75%), peritonitis (70%), bacterial overgrowth (70%) and reparative process (70%). 16 Pneumatosis intestinalis is present in only 50% of pathology specimens.

Coagulation necrosis is the predominant lesion in most specimens. 1, 16 Coagulation necrosis, when present with inflammatory lesions, is often observed in an alternating manner in adjacent microscopic fields. Inflammatory changes include acute and chronic signs of inflammation, including serositis, peritonitis, inflammatory pseudomembranes, crypt abscesses, and bacterial or fungal overgrowth. When coagulation necrosis is present with signs of inflammation, the coagulative process is often the dominant lesion. Resected tissue also demonstrates extensive apoptosis, enterocyte expression of the inducible isozyme of nitric oxide synthase (iNOS), as well as increased tissue transcripts for tumor necrosis factor α, interleukins 8 and 11, and increased messenger RNA for stromelysin 1, a matrix metalloproteinase that can degrade the extracellular matrix. [17] [18] [19] Large-vessel thrombi are unusual autopsy findings in the mesenteric arteriole system of patients with NEC. 1, 16 Small-vessel thrombi are present in 30%, but may represent secondary phenomena or autopsy artifacts.

Interestingly, reparative processes of both acute and chronic types are noted in over 50% of cases. 16 Such reparative changes include focal epithelial regeneration, granulation tissue formation and fibrosis. The latter two processes are usually noted in the mucosal and submucosal layers, but occasionally extend into the muscularis. Such fibrotic processes may later be associated with development of strictures in recovering infants. Early theories of the pathogenesis of NEC proposed a relationship between gastrointestinal ischemia, enteral alimentation and micro-organisms. 1 Although the precise contributions of these variables remain ill defined, a multifactorial pathogenesis involving these risk factors plus immature gastrointestinal function and immature host defense mechanisms seems possible for the initiation and subsequent propagation of NEC (Figs 43.2 and 43.3). 1, 12 Hypoxic-ischemic injury Potential alterations of mesenteric blood flow may occur at the large-vessel arterial level and may be associated with redistribution of systemic blood flow or may be within the gastrointestinal system itself with redistribution of local mucosal blood flow. 1, [20] [21] [22] [23] The possible systemic alterations may be associated with global hypoxia, asphyxia, exchange transfusion, arterial runoff lesions such as patent ductus arteriosus, shock or anemia. Local alterations of the mucosal circulation may be associated with less severe perturbations of these systemic factors plus those local changes occurring during enteral alimentation, polycythemia, intestinal distention or luminal exposure to bacterial toxins or inflammatory mediators. Potential events may be initiated during the prenatal period (cocaine exposure, placental insufficiency with intrauterine growth retardation) or may occur after birth. 20-23 Infants exposed to cocaine prenatally may be at higher risk for the development of NEC. Although they tend to have higher birthweights and greater gestational age than others at risk for NEC, these infants may develop NEC sooner and have fewer traditionally identifiable risk factors. Additional evidence linking prenatal intestinal ischemia and NEC among intrauterine growth-retarded infants may be demonstrated by absent or reversed diastolic blood flow detected by fetal Doppler ultrasonography. 23 The regulation of postnatal mesenteric blood flow is complex. Autoregulation of regional blood flow is determined by responses to luminal nutrients and autocrine effects of locally produced gut hormones. 1, 12 These effects may be modified by disease states, medications, the autonomic and central nervous systems, and systemic processes producing vasoconstriction or hypotensive reduction of local tissue blood flow.

Regional intestinal hyperemia in response to local enteral nutrients is associated with increased local mucosal blood flow, oxygen delivery, oxygen consumption and oxygen extraction. Oxygen-derived metabolic processes support aerobic energy-requiring intestinal mechanisms such as mucosal active transport, secretion, motility, digestion, macromolecule synthesis and cell growth. 1, 20 The postprandial intestinal mucosal hyperemia of adults is due to selective vasodilation of the mesenteric artery and is regulated by central and local mechanisms (local hormones, nutrients). It is possible that an imbalance between oxygen delivery and oxygen consumption, together with inability to augment oxygen extraction (flow-dependent model), predisposes the immature intestine to hypoxic mucosal injury during alimentation. Such hypoxic mucosal injury may alter energy-dependent processes, produce malabsorption, mucosal ulceration and ileus, increase mucosal permeability and predispose to secondary bacterial invasion.

Experiments in mature and immature models support each of these hypotheses as possible mechanisms for Pathogenesis 693

Hypertonic formula or medication Malabsorption, gaseous distention H 2 gas production by nonpathogenic bacteria Endotoxin production by nonpathogenic bacteria 

Because 95% of infants with NEC have been fed enterally before its onset, enteral alimentation has been proposed as a contributing factor in NEC. 12, 14 Various hypotheses propose that it is the composition of the milk, the rate of milk volume increments, the immaturity of gastrointestinal motility, absorptive or host defense processes, or other variables (high luminal osmolality) that contribute to the pathogenesis of NEC. 1, 12, 14 Human milk reduces the risk for NEC. 24 Animal models support a role for the breast milk macrophage, but human data suggest that anti-inflammatory cytokines and enzymes as well as immunoglobulins, specifically IgA, may have a protective advantage. Indeed, both enterally administered serum-derived IgA and human milk may reduce the incidence of NEC. 25 Human milk may also reduce allergic reactions and facilitate the development of a favorable intestinal bacterial flora, while enhancing digestion and absorption of normal nutrients.

The volume of milk fed to infants may also predispose the patient to NEC. 14,15,26-28 Excessively rapid increments of milk feeding may overcome the infant's intestinal absorptive capability, especially in the presence of altered motility, resulting in malabsorption. Malabsorbed carbohydrates contribute to enhanced intestinal bacterial gas production, resulting in abdominal distention. 1, 12, 14 High intraluminal pressure from gaseous distention may reduce mucosal blood flow, producing secondary intestinal ischemia. In addition, dissection of bacterial gas products from the intestinal lumen may produce pneumatosis intestinalis or, if gas enters the portal venous system, hepatic venous gas may be evident. Analysis of gas from the intestinal lumen and cysts of pneumatosis intestinalis reveals a profile typical of intestinal bacterial fermentation of malabsorbed carbohydrates (e.g. hydrogen, methane, carbon dioxide). 1 Earlier hypotheses had suggested that giving patients at risk nothing by mouth might reduce the incidence of NEC. Delayed feeding for asphyxiated infants or those with umbilical arterial or venous catheters and respiratory distress syndrome has not reduced the incidence of NEC. 1, 12, 14 In fact, delayed feeding may do more harm than good by increasing the risk of intestinal mucosal atrophy, cholestatic jaundice, osteopenia of prematurity and hyperalimentation-related complications. 14 Large-volume milk feedings, increased too rapidly during the feeding schedule, may place undue stress on a previously injured or immature intestine. Feeding increments in excess of 20-30 ml/kg per 24 h were associated with an increased risk of NEC in at least two studies. 12, 14 Two other studies have demonstrated the safety of 30-35 ml/kg per 24 h feeding increments. 26, 29 Such information should temper enthusiasm for excessively rapid feeding protocols for low-birthweight infants. It suggests that daily increments should be based on the clinical examination, evidence of feeding intolerance and a recommended volume increment of 20-35 ml/kg per 24 h.

Hypertonic formula and enteric medications may have direct adverse effects on mucosal blood flow and intestinal motility. Subsequent injury may predispose to NEC. 1, 12, 14 Alternatively, direct pharmacologic effects of an agent on systemic host defense (vitamin E), motility (morphine) or regional blood flow (indomethacin) may result in mucosal injury, increasing the risks for NEC in susceptible neonates. H 2 -blocking agents may decrease the risk of NEC.

There are multiple epidemiologic investigations that have provided circumstantial or direct evidence to suggest that NEC is associated with one or more microbiologic agents. 1,30 NEC has been reported to occur in epidemics or clustered episodes due to an identifiable enteric pathogen; more often no identifiable agent is discovered. 1, 30 When no agent is identifiable, this may be due to the unculturable nature of the pathogen. The suspicion of a transmissible agent is corroborated by the observation that epidemics abate following the institution or reinforcement of specific infectious disease-control measures (gowning, gloving, nurse cohorts and, especially, careful hand-washing). 1, 30 Epidemics have been associated with the recovery of no specific agent, or with the recovery of a single pathogen such as Escherichia coli, Klebsiella, Salmonella, Staphylococcus epidermidis, Clostridium butyricum, coronavirus, rotavirus and enteroviruses. 1, 31 An outbreak of NEC has been associated with Enterobacter sakazakii-contaminated powdered milk formula. 31 Additional evidence suggesting that NEC is due to an infectious agent includes the observation of related infantile diarrheal illnesses within a community or among personnel in the neonatal intensive care unit. 1 Furthermore, there are similarities between NEC (pathology, symptoms, immature susceptibility) and many enterotoxemias of young animals and humans. Such enteric toxin-mediated illnesses may be due to Clostridium, S. epidermidis or other toxin-producing enteric pathogens. [32] [33] [34] Alternatively, endotoxin production by the `normal' Gram-negative enteric flora during enteral alimentation may predispose the immature intestine to mucosal injury if endotoxin production exceeds elimination. 35 In addition transcytosis (crossing of epithelial cells by E. coli, etc.) could initiate this process; E. coli from patients with NEC has this capacity in animal models. 36 Endotoxin stimulates host inflammatory cells to produce various mediators such as tumor necrosis factor and platelet-activating factor. 35 Both of these and other inflammatory cytokines can initiate or propagate the pathologic process characterized by coagulation necrosis, inflammation, increased vascular permeability, edema, hemorrhage, local thrombosis and platelet consumption. Indeed, the immature enterocyte (epithelial cell) tends to react with excessive local production of proinflammatory cytokines when stimulated with endotoxin or interleukin 1β. The resulting cytokine mediator-induced thrombocytopenia, neutropenia, hypotensive-hypovolemic shock (third-space fluid losses), metabolic acidosis and hemorrhagic diarrhea are quite similar to the clinical manifestations of NEC in human neonates. 37 The blood culture is positive in 20-30% of patients with NEC. 1, 4 Reports of the responsible bacteremic pathogens before 1980 demonstrated a predominance of E. coli and Klebsiella. 1 Current reports of agents producing bacteremia in patients with NEC suggest that S. epidermidis is another common blood isolate. 4 It remains to be determined whether the organisms recovered in blood or peritoneal cultures are the primary pathogens or secondary invading organisms that gain access to the circulation or peritoneum through a compromised intestinal mucosa.

The predominant risk factor for NEC is prematurity, not the associated diseases of premature infants. Nonetheless, multiple potentially adverse events may produce mucosal injury, the net result being manifest as NEC. Figure 43 .2 provides potential initiating events, and Figure 43 .3 identifies additional pathologic factors that may propagate NEC once mucosal injury exceeds the immature host's ability to repair the process. Although it is hoped that one microbiologic agent or other process will be found to be responsible for NEC, it is more probable that NEC is a final common pathway for an immature intestinal response to injury. Indeed, NEC is a common cause of the systemic inflammatory response syndrome (SIRS) in neonates.

Early signs and symptoms of NEC are often non-specific; they include subtle signs of the 'sepsis syndrome' and more specific but equally subtle signs of gastrointestinal disease. 1 Non-specific extra-gastrointestinal manifestations include apnea, bradycardia, lethargy, temperature instability (hypothermia or the need to increase the Isolette temperature to maintain normal body temperature), cyanosis, mottling, cool extremities and acidosis. 1 More specific but not diagnostic gastrointestinal manifestations are related to ileus, third-space fluid losses, local coagulopathy and intestinal hemorrhage. Gastrointestinal signs and symptoms include abdominal distention, abdominal tenderness, emesis, sudden increased gastric residual volume, hematemesis, bright red blood from the rectum, absent bowel sounds, abdominal guarding and diarrhea. The latter is an uncommon isolated manifestation of NEC.

Monitoring pre-feed gastric residuals (gastric aspirates) is not helpful in predicting NEC as many unaffected neonates have gastric residuals. 38 A sudden change in the volume of gastric residuals plus abdominal distention is more ominous for NEC. Monitoring for subtle signs of gastrointestinal bleeding with stool guaiac testing is also not helpful as many unaffected premature infants have stools with occult blood, and patients with NEC may have stools negative for occult blood. Gross blood in the stool is more suggestive of NEC but is not diagnostic.

As the disease progresses in severity, there is disseminated intravascular coagulation, hypotensive (septic and hypovolemic) shock, ascites, peritonitis and intestinal perforation. Focal findings may include erythematous streaking of the anterior abdominal wall around the umbilicus and the course of the subcutaneous umbilical vein, and erythema with a mass in the right lower quadrant, representing a local perforation, with matted bowel forming a local abscess.

Disease stage should be classified as noted in 

The diagnosis of NEC is confirmed by the radiographic presence of pneumatosis intestinalis or hepatic venous gas (Figs 43.4 and 43.5) . Gastrointestinal perforation (pneumoperitoneum; Figs 43.6 and 43.7) is strong evidence for NEC, but the diagnosis must then be confirmed by histopathologic evidence.

Hepatic venous gas and ascites may also be demonstrated by abdominal ultrasonography or magnetic resonance imaging, and inapparent pneumatosis intestinalis may become more evident by performing a contrast enema. 1, 41, 42 Nonetheless, except under unusual circumstances (to rule out volvulus), contrast studies are not needed to determine the diagnosis of NEC. Ancillary laboratory evaluations may reveal thrombocytopenia with or without evidence of disseminated intravascular coagulation, anemia, neutropenia, metabolic acidosis from septic or hypovolemic shock, respiratory acidosis from increased intra-abdominal pressure and poor diaphragm excursion, increased breath hydrogen excretion, raised fecal calprotectin levels and radiographic signs of ileus, an isolated dilated intestinal loop or ascites. 1, 43 The differential diagnosis of NEC is outlined in Table 43 .2. Idiopathic, focal, spontaneous, isolated intestinal perforation is also a common cause of pneumoperitoneum in the pre-term and, less often, full-term neonate. The onset is usually sudden and occurs earlier in life than that of NEC. Prior feeding may not be present and pneumatosis intestinalis is absent. Affected patients usually have respiratory distress syndrome and a lower birthweight and gestational age than patients with NEC. The abdominal wall may appear blue, and radiography shows free air and a gasless abdomen. Overall survival is better than patients with NEC who develop pneumoperitoneum. 44, 45 PREVENTION NEC may not be completely eliminated, but certain interventions have been demonstrated to lower the incidence. Oral administration of an IgA-IgG preparation has been demonstrated to reduce the incidence of NEC; however, intravenous or oral administration of IgG has not affected the incidence. 46 Human milk significantly reduces the incidence of NEC among pre-term infants fed donor or their mother's milk. 14, 47 Prenatal administration of corticosteroids lowers the incidence of NEC. 48 Steroids enhance lung maturation and are thought to accelerate intestinal maturation, thus potentially reducing the 'immaturity' factor in the pathogenesis of NEC. Postnatal steroids may or may not affect the incidence of NEC.

Judicious slow enteral feeding protocols (no volume increments exceeding 20-35 ml/kg per 24 h) may reduce the occurrence of NEC. 14, 27 There is sufficient circumstantial evidence from multiple case-control studies to suggest that moderately slow feeding protocols are associated with a lower incidence of NEC. Initiation of 10 days of small-volume feedings (20 ml/kg daily) (gut stimulation, minimal enteric feeds) has been demonstrated to reduce the incidence of NEC when compared to premature infants given advancing feeding (increments of 20 ml/kg daily). 27 In addition, arginine supplementation 49 may reduce the incidence of NEC, whereas enteral antibiotics (vancomycin or kanamycin) reduce the incidence of NEC but also increases the incidence of microbial flora resistance to the chosen antibodies and thus cannot be recommended routinely. 50, 51 Finally, prevention of premature breath, for instance by administering 17α-hydroxyprogesterone caproate to the mother, also reduces the incidence of NEC. 52 

NEC has a wide spectrum of severity. The mildest form manifests as hemorrhagic colitis with or without pneumatosis coli; the more fulminant state is similar to that noted in patients with Gram-negative septic shock, commonly referred to as SIRS.

Abdominal distention is a universal feature of NEC. Significant abdominal distention may reduce the mesenteric arterial perfusion pressure, thus exacerbating a previously compromised intestinal blood flow. Models of the effects of increased intra-abdominal pressure have reported such adverse consequences as increased systemic vascular resistance, decreased cardiac output, decreased urine output and 'apparent' hypovolemia. 53, 54 Surgical decompression of increased intra-abdominal pressure in human adults restores systemic arterial oxygenation, cardiac output and urine production within 15 min of the procedure. 53 Increased intra-abdominal pressure in patients with NEC is due to the development of tense ascites, marked intestinal gas production, stasis (ileus) and inflammatory fluid exudation with hemorrhage into the lumen of the small and large intestines. It is imperative to reduce abdominal distention with nasogastric tube placement and no further formula feeding (NPO -nil per os). The decompression tube should be the largest that the patient can tolerate. Paracentesis with placement of an intra-abdominal drain under local anesthesia has been helpful in stage II or III disease. Finally, if the patient fails to respond to medical management or abdominal drain placement within 24-48 h of the onset of illness, exploratory laparotomy can result in abdominal decompression by removal of necrotic tissue and inflammatory exudate (Table 43. 3).

The associated bacteremia in approximately 20-30% of patients with NEC is probably not the primary cause of the disease. Nonetheless, appropriate antimicrobial therapy must be directed against these bacteria, even when the bacteremia is due to bowel injury and secondary bacterial invasion. Patients with both NEC and bacteremia usually have more severe disease and a higher mortality rate. 1 Although the precise antibiotic regimen for the treatment of NEC has not been determined, the clinician must remain flexible, as there are changes in the pathogens recovered in patients with NEC, and each neonatal intensive care unit has different bacteria with different antimicrobial resistent patterns. 1, 4, 30, 50 This may reflect a bacterial shift in the fecal colonization of premature infants. Nonetheless, the changing pattern of the agents recovered during bacteremia requires close scrutiny and appropriate modification of antimicrobial therapy. Traditional antimicrobial treatment of NEC employs systemic administration of a semisynthetic penicillin (ampicillin, ticarcillin) and an aminoglycoside (gentamicin, kanamycin). Evidence suggests a beneficial response with the use of vancomycin and cefotaxime. 55 Many recommend anaerobic coverage with clindamycin or metronidazole. Vigilant attention to the microbiology of blood, fecal and peritoneal cultures in patients with NEC is needed for appropriate modification of antibiotic therapy (see Table 43 .1 for duration of antimicrobial therapy) 56,57 . Perforation with subsequent bacterial peritonitis is managed with abdominal drain placement or surgery. The treatment of severe NEC manifested as SIRS is not unlike that of other causes of bacteremia-associated hypotension. Endotoxin-stimulated production of inflammatory mediators such as bradykinin, tumor necrosis factor or platelet-activating factor results in increased vascular permeability, large transcapillary fluid loss, increased pulmonary artery pressure with hypoxia, lactic acidosis and hypotension. Hypermetabolism increases oxygen requirements. Once stabilized from the septic shock state, all patients require parenteral nutrition while NPO. Fluid losses plus the initial vasodilation increase water and electrolyte requirements. Despite rapid killing of bacteria by antibiotics, the diverse effects of bacterial products (endotoxin, etc.) are still evident. The net result is a markedly reduced cardiopulmonary ability to meet the oxygen requirements of the peripheral tissues. [58] [59] [60] The decreased cardiac output in septic shock (and NEC) may be due to pooling of blood and fluid in the capacitance peripheral vessels and loss of fluid in third spaces as well as a specific myocardial dysfunction characteristic of severe bacteremic states. The myocardial depression is not due to ischemia but rather to various mediators released as a response to inflammation and hypotension. Circulatory failure results in a flow-limited ability to provide the oxygen necessary to support energy metabolism (local tissue oxygen consumption). Circulatory failure causes tissue hypoxia and metabolic (lactic) acidosis. Some of the cellular defects of oxygen utilization may not be due to hypoxia-ischemia alone. Tumor necrosis factor produced during endotoxemia has various metabolic consequences that may interfere with mitochondrial oxygen utilization. 58 Methods used to treat septic shock must take into consideration the linear relationship between oxygen delivery and local tissue oxygen consumption. 58 This flow-dependent relationship can be improved by restoring the circulating blood volume (preload) with fluid resuscitation and by improving myocardial contractility with inotropic sympathomimetic agents.

An acute 'adult' respiratory distress-like syndrome (ARDS) may be observed in patients with NEC. This is due in part to inflammatory or vasoactive mediators producing non-cardiogenic pulmonary edema. Hypoxia is exacerbated by increased pulmonary artery pressure, abdominal distention with reduced diaphragmatic excursion, and myocardial contractile failure. Adequate oxygen delivery is closely dependent on appropriate ventilator management of patients with NEC. Methods that improve oxygen delivery also reduce local lactate production and improve metabolic acidosis. A successful outcome in patients with septic shock is related to the ability of the therapeutic measures to improve cardiac output. In addition to support for the failing circulation, careful attention must be given to the pulmonary problems associated with NEC.

The application of these principles to the therapy of NEC must emphasize aggressive use of fluid resuscitation. If fluid administration is unsuccessful in restoring perfusion and urine production, or in correcting the metabolic acidosis, inotropic drugs (dopamine, dobutamine) can be used to improve oxygen delivery by improving myocardial contractility and occasionally by vasodilation. The administration of inotropic, vasodilator or vasopressor agents must be carefully titrated against peripheral perfusion, blood pressure, urine production, metabolic acidosis and central venous pressure if available. The therapeutic balance between vasopressor (myocardial contractility) agents and vasodilation (afterload reduction) is often difficult to achieve but may benefit from additional fluid therapy.

Abdominal drainage by percutaneous drain placement may acutely decompress the abdomen and improve many of the adverse cardiopulmonary complications of NEC due to a tense abdomen and compromised diaphragms. It may be particularly helpful in improving oxygenation and ventilation. 61 Intestinal perforation is a traditional indication for exploratory laparotomy. In certain high-risk, unstable patients (often weighing less than 1000 g with stage IIb or III NEC), percutaneous placement of an abdominal drain under local anesthesia has been performed. [61] [62] [63] [64] In many but not all patients with NEC managed with paracentesis and a drain, exploratory laparotomy is necessary 24-72 h later. Drain placement is even more beneficial for patients with isolated idiopathic intestinal perforations. Fewer patients with isolated perforations may require immediate exploratory laparotomy in 24-48 h. Additional indications for surgery include progressive clinical deterioration despite aggressive medical management (see Table 43 .3), and in the convalescent stages for the resection of strictures and enteric fistulae. All patients treated with abdominal drain placement must be considered at risk for stricture formation. If percutaneous drainage is used as the first surgical treatment, exploratory laparatomy is then indicated for deterioration, signs of intestinal obstruction or failure of a second drain to relieve pneumoperitoneum.

Surgical management should attempt to preserve as much viable bowel as possible, resecting only the most obviously necrotic and gangrenous tissue. In circumstances of NEC totalis, a high diverting jejunostomy is recommended. 65 A second laparotomy is performed within 48-72 h if the patient remains critically ill to determine whether what previously looked like non-viable tissue was actually viable. 65 This approach may avoid massive resection and subsequent development of the short gut syndrome. 66 For patients with minimal and well defined disease, some surgeons recommend that primary anastomosis be performed after resection of dead bowel, at the time of the initial laparotomy. Such patients should be observed carefully for development of strictures (at the anastomosis and other sites) or recurrent NEC. Strictures may present with signs of obstruction (emesis, obstipation, abdominal distention), sepsis or gastrointestinal bleeding.

Neonatal necrotizing enterocolitis: pathogenesis, classification and spectrum of disease

A regional study of underlying congenital diseases in term neonates with necrotizing enterocolitis

Necrotizing enterocolitis mortality in the United States, 1979-85

Outcome of neonatal necrotising enterocolitis: results of the BAPM/CDSC surveillance study

Impact of necrotizing enterocolitis on length of stay and hospital charges in very low birth weight infants

National Nosocomial Infections Surveillance System. Nosocomial infections among neonates in high-risk nurseries in the United States

Nutritional and other postoperative management of neonates with short bowel syndrome correlates with clinical outcomes

Gastrointestinal perforation and peritonitis in infants and children: experience with 179 cases over ten years

Part 1: Current controversies in the understanding of necrotizing enterocolitis

Epidemiology of neonatal necrotizing enterocolitis: a population-based study

Frequency of and risk factors for necrotizing enterocolitis in infants born before 33 weeks of gestation

Neonatal necrotizing enterocolitis: bridging the basic science with the clinical disease

Necrotizing enterocolitis in neonates with congenital heart disease: risk factors and outcomes

The relationship of neonatal feeding practices and the pathogenesis and prevention of necrotizing enterocolitis

Rapid versus slow rate of advancement of feedings for promoting growth and preventing necrotizing enterocolitis in parenerally fed lowbirth-weight infants

Pathology of neonatal necrotizing enterocolitis: a ten-year experience

Role of platelet-activating factor in the pathophysiology of necrotizing enterocolitis

Intestinal cytokine gene expression in infants with acute necrotizing enterocolitis: interleukin-11 mRNA expression inversely correlates with extent of disease

Matrix metalloproteinases in necrotizing enterocolitis

Intestinal ischemia and necrotizing enterocolitis

Gut blood flow velocities in the newborn: effects of patent ductus arteriosus and parenteral indomethacin

Time of onset of necrotizing enterocolitis in newborn infants with known prenatal cocaine exposure

Ultrasound predictors of neonatal outcome in intrauterine growth restriction

Donor human milk versus formula for preventing necrotizing enterocolitis in preterm infants: systematic review

Prevention of necrotizing enterocolitis in low-birth-weight infants by IgA -IgG feeding

Randomized trial of 'slow' versus 'fast' feed advancements on the incidence of necrotizing enterocolitis in very low birth weight infants

Prolonging small feeding volumes early in life decreases the incidence of necrotizing enterocolitis in very low birth weight infants

The incidence of necrotizing enterocolitis after introducing standardized feeding schedules for infants between 1250 and 2500 grams and less than 35 weeks of gestation

The effect of feeding volume on the clinical outcome in premature infants

Nosocomial necrotizing enterocolitis outbreaks: epidemiology and control measures

Outbreak of necrotizing enterocolitis associated with Enterobacter sakazakii in powdered milk formula

Role of bacterial toxins in neonatal necrotizing enterocolitis

Analysis of enteric coagulase-negative staphylococci from neonates with necrotizing enterocolitis

Effect of lipopolysaccharide on intestinal intramucosal hydrogen ion concentration in pigs: evidence of gut ischemia in a normodynamic model of septic shock

Bacterial toxins and enteral feeding of premature infants at risk for necrotizing enterocolitis

Escherichia coli transcytosis in a caco-2 cell model: implications in neonatal necrotizing enterocolitis

Inflammation in the developing human intestine: a possible pathophysiologic contribution to necrotizing enterocolitis

A critical analysis of the routine testing of newborn stools for occult blood and reducing substances

Colonic pneumatosis intestinalis in preterm infants: different to necrotizing enterocolitis with a more benign course?

Pneumatosis coli -a benign form of necrotizing enterocolitis

Acidosis and hepatic portal venous gas: indications for surgery in necrotizing enterocolitis

Sonographic findings in necrotizing enterocolitis with paucity of abdominal gas as the initial symptom

Faecal calprotectin concentrations and diagnosis of necrotizing entercolitis

A comparison of the clinical presentation and outcome of focal intestinal perforation and necrotizing entercolitis in very-low-birthweight neonates

Spontaneous localized intestinal perforation in very-low-birth-weight infants: a distinct clinical entity different from necrotizing enterocolitis

Oral immunoglobulin for preventing necrotizing enterocolitis in preterm and low birth-weight neonates

Breast milk and necrotising enterocolitis

Prenatal and postnatal corticosteroid therapy to prevent neonatal necrotizing enterocolitis: a controlled trial

Arginine supplementation prevents necrotizing enterocolitis in the premature infant

Double blind, randomized, placebo controlled study of oral vancomycin in prevention of necrotizing enterocolitis in preterm, very low birthweight infants

Enteral antibiotics for preventing necrotizing enterocolitis in low birthweight or preterm infants

Prevention of recurrent preterm delivery by 17alpha-hydroxyprogesterone caproate

Cardiovascular, pulmonary, and renal effects of massively increased intraabdominal pressure in critically ill patients

Acute renal failure associated with increased intra-abdominal pressure

Comparison of two antibiotic regimens for neonatal necrotizing enterocolitis

Intraabdominal infections

Emergence of antibiotic-resistant bacteria in cases of peritonitis after intraabdominal surgery affects the efficacy of empirical antimicrobial therapy

Cellular oxygen metabolism during sepsis and shock: the relationship of oxygen consumption to oxygen delivery

Role of endotoxemia in cardiovascular dysfunction and mortality: Escherichia coli and Staphylococcus aureus challenges in a canine model of human septic shock

Sequential physiologic interactions in pediatric cardiogenic and septic shock

Primary peritoneal drainage for increasing ventilatory requirements in critically ill neonates with necrotizing enterocolitis

Controversies in the management of necrotizing enterocolitis

Peritoneal drainage as primary management in necrotizing enterocolitis: a prospective study

A meta-analysis of peritoneal drainage versus laparotomy for perforated necrotizing enterocolitis

Is there a role for high jejunostomy in the management of severe necrotizing enterocolitis?

Complications after surgical intervention for necrotizing enterocolitis: a multicenter review

Gellis & Kagan's Current Pediatric Therapy, 15th edn