Neonatal Gastrointestinal Bleeding: Update

Gastrointestinal bleeding (HD) is a relatively uncommon presentation in the NICU.1 Typically seen in sick, premature neonates, HD is most commonly classified based on the presumed location of blood loss: upper or upper (i.e., above the ligament of Treitz) or inferior or low (i.e., below the ligament of Treitz).2 Although the absolute blood volume of neonates is significantly less than that of older patients, rapid blood loss is more compensated easily, resulting in a significantly lower HD mortality rate in infants and children.3

HD in neonates is characterized by many phenotypes that exist in a wide clinical spectrum of lesser or greater severity. Few epidemiological studies have measured the frequency of HD in infants and young children. Most of these analyzes are retrospective, and the only prospective cohort showed an HD rate of 6.4% in children admitted to the pediatric NICU as described by Lacroix et al.4

Table 1 describes the most common disorders causing superior or inferior HD in neonates. When considering potential etiologies of HD in a newborn, the medical team should seek to establish whether this finding represents true hemorrhage (e.g., esophagitis, variceal bleeding, gastritis, cow’s milk protein intolerance [LVPI], colitis) or a “clinical imitation” of hemorrhage. Specific mimics of HD that are common in neonates include maternal blood ingested during the birth process or blood ingested from the mother’s bleeding nipple during direct breastfeeding.5 When clinical mimics of HD have been excluded, the various phenotypes of HD must be considered. true hemorrhage.

> HD higher or high

Hematemesis is the most common clinical sign of high HD in the newborn.

Hematemesis may result from esophagitis, gastritis, or enteritis. The most common presentation of clinically significant hematemesis requiring urgent treatment is a stressful delivery.6 Other causes of acute high HD include sepsis, coagulopathy due to vitamin K deficiency, iatrogenic injury to the gastrointestinal mucosa from feeding tube placement, secondary variceal bleeding to portal hypertension due to liver failure, gastroesophageal reflux disease (GER), gastritis/peptic ulcer, infection, and exposure to pharmacotherapy (both in utero and postnatally). Certain common clinical scenarios in the NICU appear to predispose neonates to high HD, such as medical treatment for a clinically significant patent ductus arteriosus with ibuprofen (8.9 neonates with secondary HD per 100) or the use of systemic corticosteroids (e.g. , dexamethasone).7,8

There are also many chronic and insidious causes of high HD in neonates. Specifically, neonates may undergo a sensitization phase following the introduction of enteral nutrition during exposure to casein and whey proteins, which may lead to IPLV.9 This diagnosis may be seen as gastroesophageal reflux in the setting of a poor growth and may also be accompanied by clinical signs of low HD, such as hematochezia (food protein-induced proctocolitis).

Although IPLV is a known enhancer of GER in newborns, GER disease can also present as HD separately from IPLV. This usually occurs in neonates with other specific comorbidities, such as neuromuscular diseases, congenital anatomic anomalies (eg, hiatal hernia, esophageal duplication, gastric outlet obstruction, intestinal atresia), or enteric nervous system dysfunction. In other scenarios, high HD may be the result of an infection (e.g., Candida albicans , herpes simplex virus, cytomegalovirus) secondary to relative immunosuppression in the sick preterm neonate.

Rare causes of superior HD in newborns include congenital dysfunction of the coagulation cascade (e.g., factor [F] II, FVII, FVIII, FIX, FX, or FXIII deficiency, presence of inhibitors, or acquired) and vascular anomalies ( e.g., gastrointestinal hemangioma, Prune Belly syndrome/Eagle-Barrett syndrome, Dieulafoy lesion, Kasabach-Merritt syndrome or gastrointestinal lesions due to hereditary hemorrhagic telangiectasia).3,10,11

> Lower or low GHD

Unlike high HD, low HD is more commonly associated with hematochezia.

As detailed above, the most common presentation of hematochezia in the newborn is due to IPVL. Other common associations with low HD in decreasing order of incidence include necrotizing enterocolitis (NEC), coagulopathy, infectious colitis, congenital enterocolonic malformation, enterocolitis associated with Hirschsprung disease, and intestinal vascular lesions (e.g., hemangiomas and vascular malformations).

Vascular malformations and anorectal disorders in neonates have a large number of etiologies that predispose to bleeding, including anorectal malformations, anal fissures, and hemorrhoids. Malignant masses or intra-abdominal congenital embryonic malformations can compress the abdominal vasculature and manifest as external hemorrhoids. In particular, neonatal hemorrhoids can be evolutionary and temporary, secondary to prolonged pressure on the abdominal cavity of the newborn during childbirth. Historically they have been a cause of concern for chronic liver disease. In fact, up to 35% of children with portal hypertension will develop a hemorrhoid, although no cases have been reported in neonates.12

Neonates presenting with signs of HD should undergo a detailed history and a dedicated physical examination considering both the history and exposures of the pregnant woman, as well as the peripartum course of the newborn up to the time of HD. The laboratory and diagnostic modalities are summarized in Table 2.

In the past, it was common for NICUs to perform periodic aspiration of infants’ gastric contents before feeding in an attempt to clinically correlate the nature of the aspirate with the child’s progress on the tube feeding plan.13 There is no evidence that support regular, routine prefeeding gastric aspiration as a predictor of the presence of bleeding lesions in neonates.14 Extrapolating from small pediatric cohorts as well as studies in adults, gastric lavage is not recommended in the presence of high HD. neonates in critical conditions. The main concern is that neonates who present signs of high HD may have decreased luminal integrity and are at high risk of iatrogenic tissue injury and/or perforation during the procedure. Furthermore, the rate of false negative results with gastric lavage is high, which decreases the negative predictive value of the diagnostic procedure. Gastric lavage is also a poor predictor of high-risk lesions and provides little prognostic value in determining when endoscopy is warranted.15,16,17

For patients with “clinical mimicry” of hemorrhage, such as maternal blood ingested during childbirth or breastfeeding, the Apt test can be performed to determine whether the blood that is present in a newborn’s emesis or stool is of maternal or if it is due to the patient’s own bleeding. When the blood is of maternal origin, newborns generally appear well and have normal vital signs and blood counts.3,5

Laboratory evaluation of neonates with HD should be tailored to the child’s presentation. A serum analysis in the setting of clinical decompensation should include a complete blood count with differential count, a comprehensive metabolic panel, inflammatory markers (e.g., erythrocyte sedimentation rate, C-reactive protein, procalcitonin), and coagulation studies to evaluate anemia and thrombocytopenia, electrolyte alterations, hepatitis and/or compromised liver function. Additionally, this information will help determine if a goals-based transfusion protocol is indicated.

When NEC is suspected, fecal calprotectin can provide confirmatory quantification of luminal inflammation to guide treatment. Calprotectin is a heterocomplex calcium and zinc binding protein. Calprotectin is released in large quantities after mucosal neutrophil aggregation and cell death in the intestinal lumen and is then excreted in the feces. This test is resistant to enzymatic degradation and is therefore superior to fecal leukocytes in the non-invasive measurement of intestinal and colonic inflammation. This test has demonstrated clinical utility in premature newborns with NEC.18,19,20,21

For infants with early NEC, studies have shown a significant early transient increase in fecal calprotectin compared to healthy infants.18,19,20,21 Additionally, elevation of fecal calprotectin can be detected in very low birth weight infants. born 12 to 48 hours before the onset of external clinical signs of moderate NEC.18,19,20,21 In predominantly European studies, mean fecal calprotectin levels in infants with NEC were between 210 and 400 mg/g of stool. 18

Fecal calprotectin has demonstrated a sensitivity of 86% and a specificity of 93% when applied to infants at increased risk for NEC.18,19,20,21 The drawback of this test is that most centers will have to refer the sample, and it may take 3 to 5 days to obtain the results. Fortunately, rapid point-of-care fecal calprotectin analyzers have been developed, demonstrating strong comparability with traditional calprotectin analyzers.22,23 This may present an opportunity to improve quality of NEC care in a large NICU. .

In addition to fecal calprotectin, an additional robust stool study, gastrointestinal film array technology, is needed to diagnose infectious etiologies of HD. This film array technology uses the polymerase chain reaction to detect viral, bacterial, and parasitic pathogens in the gastrointestinal tract.24,25 Of note, many of these stool polymerase chain reaction tests include the detection of Clostridium difficile toxins A and B.

It is important to recognize that despite the relative state of immunosuppression that many critically ill neonates possess, testing positive for C. difficile species and/or toxin A/B is unlikely to provide clinically meaningful information. In cases where a neonate tests positive for C. difficile , the child is likely colonized with the bacteria.

It is unlikely that newborns possess the colonic receptors that would facilitate toxin trafficking and the subsequent inflammatory response, which would lead to C. difficile colitis . Guidelines and expert opinions from professional societies, including the American Society for Health Epidemiology, the Infectious Diseases Society of America, and the American Academy of Pediatrics, do not recommend routine testing for C. difficile in newborns.26

Additionally, the authors recommend that clinical teams seek alternative diagnoses if they are considering C. difficile colitis in a critically ill neonate with hematochezia and diarrhea. It is also worth mentioning that film array technology can detect pathogens shed in feces that are causing disease in the upper gastrointestinal tract. This may present a useful opportunity to perform diagnostic testing in specific scenarios where infectious esophagitis/gastroenteritis is suspected but upper endoscopy is contraindicated or not warranted.

Diagnostic radiology has demonstrated an evolving role in both the research and treatment of HD in infants.27

Bedside x-rays may be useful to rule out intestinal perforation or obstruction. Fluoroscopy can also be used to evaluate upper gastrointestinal strictures or obstructive pathologies that may be the primary cause of HD (eg, esophageal membrane, malrotation with volvulus, gastric outlet obstruction).

Barium enemas help comment on the caliber and presence of a transition zone in children with low HD suspected secondary to Hirschsprung disease. Radiography was traditionally used in the categorization of the Vermont Oxford Network’s staging criteria for NEC (ie, the modified Bell classification). Evidence is emerging to support the utility and specificity of bedside/point-of-care ultrasound in these settings, especially when abdominal radiograph findings are inconclusive.28

Specifically, ultrasound has the benefit of the absence of radiation and the ability to provide a dynamic transmural assessment of the intestine in real time. Consequently, abdominal ultrasound can provide early detection of intestinal necrosis before the development of intestinal perforation that can be seen on radiography.29

Neonatal abdominal ultrasound is best performed using high-resolution transducers with a frequency of 8 to 15 MHz to visualize the intestinal wall. This modality has proven useful in the detection of NEC when the Faingold protocol is used to evaluate each quadrant of the abdomen in both the transverse and sagittal planes.30

In some cases, it may be difficult to identify the source of HD. In these rare cases where HD may occur in the jejunum or ileum, the small size of the newborn precludes more advanced testing, such as capsule video endoscopy, that might otherwise be offered to children or adults. Therefore, more sophisticated computed tomography (CT) or nuclear medicine testing may be indicated. CT angiography is readily available, but is associated with a sensitivity as low as 40%.2

An additional drawback of CT angiography in neonates with dark HD is that it can only detect active bleeding at 1 to 2 ml/min.2 However, angiography offers the advantage of its diagnostic and therapeutic potential through direct embolization , especially when massive bleeding or large vascular abnormalities preclude the use of endoscopy. In contrast, scintigraphy with technetium-99 labeled red blood cells can help detect slow bleeding at rates as low as 0.1 ml/min.31

The disadvantage of such nuclear medicine evaluation is that, although it can detect hemorrhage, specific localization of an obscure HD may be difficult. In particular, the labeled RBC scan can be difficult to interpret in a small patient, such as a newborn, especially in the setting of sepsis and significant inflammation in which the radiotracer can potentially be distributed diffusely throughout the body.32

> Medical management

The central principle in the treatment of children with HD is to adequately address the airway, breathing, and circulation. Often, children with a single episode of hematemesis, ongoing GER, and poor growth will meet criteria for IPLV. In this case, it is recommended to follow the Clinical Practice Guidelines of the North American Society of Pediatric Gastroenterology, Hepatology and Nutrition for the evaluation and treatment of reflux.33 Rosen et al. demonstrated that in the management of childhood reflux, amino acid/elemental or extensively hydrolyzed formulas are the medical therapy of choice for populations at risk with GER as opposed to acid suppression therapy.33 It is recommended to review the guideline of the American Society of Parenteral and Enteral Nutrition to select an elemental or extensively hydrolyzed formula for these newborns. ( http://www.nutritioncare.org/Guidelines_and_ Clinical_Resources/EN_Formula_Guide/ES_Fórmulas_Infantiles/)

Meeting all dietary reference intake values ​​for neonates, these formulas provide complete nutrition and are designed to present lower molecular weight forms of casein and whey compared to the standard formula. Therefore, this nutrition is more likely to be absorbed by the neonate with less energy expenditure during the thermic effect of meals, thus improving growth.

Occasionally, NICU teams may encounter infants with scant, bright red rectal bleeding located at the anorectal edge, characterized by mucosal tearing. Anal fissures develop most commonly in newborns with constipation secondary to high calorie density feedings that do not meet their daily free water needs. In these cases, conservative management with topical zinc oxide, addressing any free water deficit, and pharmacotherapy with the laxative polyethylene glycol 3350 will resolve the anal fissure.34

Anecdotal evidence suggests that infants with constipation may be supplemented with nonabsorbable carbohydrates (e.g., sorbitol in prune, pear, and apple juice) to soften stools and facilitate defecation (typical dosage is ~2 ounces per day). However, when infants experience complications such as anal fissures due to constipation, polyethylene glycol 3350 represents an evidence-based and proven safety option to prevent stool from causing repeated trauma to anal tissue.34,35 When infants have multiple anal fissures and/or your constipation is recalcitrant to conservative treatment, manometry with or without bedside rectal suction biopsy can be performed to rule out aganglionosis of the distal enteric nervous system.36

Conversely, when neonates present with decompensation due to hypovolemia and/or clinically significant anemia, a goal-based transfusion protocol that addresses the patient’s specific coagulopathy is indicated. During acute treatment of clinically significant HD, it is prudent to initiate acid suppression therapy with intravenous proton pump inhibitors (PPIs). The use of a histamine-2 receptor antagonist has been shown to decrease the risk of HD in high-risk infants.37 However, there is no evidence to support the idea that acid suppression therapy alters mortality or the need for of blood transfusion in infants.37 From a preventive point of view, there is moderate evidence recommended by the American Gastroenterological Association for the use of specific probiotic strains in premature neonates (<37 weeks gestational age) to prevent the development of NEC .38

When performing acute resuscitation on a critically ill neonate with clinically significant HD, the available evidence suggests that medical teams should provide the patient with an early initial bolus of intravenous PPI. The intravenous PPI can then be continued every 12 hours as a scheduled infusion. Evidence suggests that this bolus approach is probably equivalent to a continuous infusion of intravenous PPI.39

When variceal bleeding is suspected, intravenous vasoactive agents that affect the splanchnic vasculature can be used.2 Specifically, the somatostatin analogue, octreotide, and the antidiuretic hormone, vasopressin, can be used safely and effectively.40,41, 42 Because vasopressin can cause significant peripheral vasoconstriction and may put the neonate at risk for acute renal failure, octreotide is usually the drug of choice in the setting of significant variceal hemorrhage.

> Endoscopic evaluation and therapeutic management 

When medical treatment is not completely effective or when HD continues long-term, consultation with a pediatric gastroenterologist should be made to consider endoscopic evaluation(s). Of note, the limited evidence available for performing endoscopy in infants with HD suggests that this procedure has limited diagnostic or therapeutic benefit.43 Bose et al. identified the source of blood loss in only 7 of 56 infants with high HD during upper endoscopy.43 Only 3 of 56 infants underwent therapeutic intervention during endoscopy.43 Five percent (n=3) of these infants presented gastrointestinal perforation in the acute postoperative period after endoscopy.43

When considering endoscopy for these neonates, medical teams should be aware of the effects of patient size and clinical status on pre-endoscopy planning and the likelihood of success. Newborns will require smaller gastroscopes (typically < 6 mm OD and 2.0–2.4 mm working channel).44 Typically, if a colonoscopy is required, the pediatric gastroenterologist will use the same gastroscope for the lower endoscopy.

Beyond the endoscope itself, members of the clinical team must recognize that not all pediatric endoscopy centers will have access to the complete and specific arsenal of hemostatic devices for neonates before the appearance of clinical signs of bleeding. Specifically, endoscopic technologies that can be used in therapeutic intervention of neonates must be able to pass through a 2.0 to 2.4 mm working channel.

Typically, these tools include an injection needle to provide submucosal epinephrine and sclerosants, a bipolar probe for contact electrocoagulation, and an argon plasma coagulation probe to provide non-contact thermal hemostasis. There are currently no hemostasis clips on the US market that can pass through a 2.0 to 2.4 mm gastroscope working channel. Additionally, modifications can be made to the hemostatic powder spray to be administered through a 5 Fr catheter, allowing passage through a 2.0 mm working channel. However, use of some of these devices may be considered "off-label" by the manufacturer and should therefore be managed at the discretion of the endoscopist.

The participation of the primary medical team in planning the endoscopic procedure is essential. Specifically, its members should assist the endoscopy team preoperatively, ensuring that the patient is instructed not to eat anything by mouth, continuing pre-endoscopy acid suppression therapy with an intravenous PPI, and considering the administration of Intravenous erythromycin approximately 30 minutes before endoscopy (in the case of high HD). This single dose of erythromycin, a motilin receptor agonist, is unlikely to increase the risk of pyloric stenosis in this population.

The use of intravenous erythromycin or metoclopramide immediately before endoscopy has been shown to improve visualization of the gastrointestinal lumen and decrease the need for repeat procedures in patients with high HD.45,46 Data from adults suggest that endoscopy is not recommended. routine screening in patients with high HD, but neonatal data are lacking.47,48,49

> Evaluation and surgical intervention

In rare cases, medical, endoscopic, or angiographic treatments will fail and newborns with HD will require laparotomy to identify the source of bleeding. Routinely, these operative interventions involve suture ligation of the culprit blood vessels. Unfortunately, there are cases where additional aggressive measures are required. Tissue resection is warranted when lesions are extensive (e.g., gastric wedge resection for Dieulafoy lesion, gastrectomy/antrectomy for severe peptic ulcers, selective splenorenal bypass for hypertensive portal gastropathy).50,51

HD in newborns includes a broad spectrum of disease morbidity, from minor symptoms of reflux and growth failure to severe, clinically significant anemia requiring intensive care.

In recent years, multiple diagnostic tools have emerged, including fecal calprotectin and bedside ultrasound, which have proven useful in the early recognition of sources of HD in neonates. More evidence continues to demonstrate that traditional medical therapy with an intravenous PPI is well tolerated and that upper endoscopy has limited diagnostic and therapeutic value but may be necessary in certain clinical scenarios.

Finally, additional research and quality improvement are underway to determine how best to prevent, recognize, and manage HD in critically ill neonates.

Gastrointestinal bleeding is most commonly classified based on the presumed location of blood loss, and in neonates it is characterized by diverse phenotypes that exist across a broad clinical spectrum of lesser or greater severity.

When considering potential etiologies of HD in a newborn, one should try to establish whether this finding represents a true hemorrhage or a “clinical imitation” of it. Currently, there are multiple diagnostic tools that will help in early recognition of the cause of hemorrhage.

Medical, endoscopic and/or surgical management will depend on the source of HD, the clinical status of the patient and their evolution throughout the process.

(See Tables below)

Table 1. Etiologies of bleeding from the upper and lower gastrointestinal tract.

Table 2. Diagnostic tools for the evaluation of HD.