The Indian Journal of Pediatrics

, Volume 80, Issue 4, pp 326–333

Approach to a Child with Upper Gastrointestinal Bleeding

Authors

    • Department of Pediatrics Advanced Pediatrics CentrePost Graduate Institute of Medical Education and Research
  • Puneet Jain
    • Department of Pediatrics Advanced Pediatrics CentrePost Graduate Institute of Medical Education and Research
  • M. Jayashree
    • Department of Pediatrics Advanced Pediatrics CentrePost Graduate Institute of Medical Education and Research
  • Sadhna Lal
    • Department of GastroenterologyPost Graduate Institute of Medical Education and Research
Symposium on PGIMER Management Protocols in Gastrointestinal Emergencies

DOI: 10.1007/s12098-013-0987-x

Cite this article as:
Singhi, S., Jain, P., Jayashree, M. et al. Indian J Pediatr (2013) 80: 326. doi:10.1007/s12098-013-0987-x

Abstract

Upper gastrointestinal bleeding (UGIB) is a potentially life threatening medical emergency requiring an appropriate diagnostic and therapeutic approach. Therefore, the primary focus in a child with UGIB is resuscitation and stabilization followed by a diagnostic evaluation. The differential diagnosis of UGIB in children is determined by age and severity of bleed. In infants and toddlers mucosal bleed (gastritis and stress ulcers) is a common cause. In children above 2 y variceal bleeding due to Extra-Hepatic Portal Venous Obstruction (EHPVO) is the commonest cause of significant UGIB in developing countries as against peptic ulcer in the developed countries. Upper gastrointestinal endoscopy is the most accurate and useful diagnostic tool to evaluate UGIB in children. Parenteral vitamin K (infants, 1–2 mg/dose; children, 5–10 mg) and parenteral Proton Pump Inhibitors (PPI’s), should be administered empirically in case of a major UGIB. Octreotide infusion is useful in control of significant UGIB due to variceal hemorrhage. A temporarily placed, Sengstaken-Blakemore tube can be life saving if pharmacologic/ endoscopic methods fail to control variceal bleeding. Therapy in patients having mucosal bleed is directed at neutralization and/or prevention of gastric acid release; High dose Proton Pump Inhibitors (PPIs, Pantoprazole) are more efficacious than H2 receptor antagonists for this purpose.

Keywords

Upper gastrointestinal bleedingEndoscopyVariceal bleedOctreotide

Introduction

Upper gastrointestinal bleeding is usually defined as bleeding from a site proximal to the ligament of Treitz at the level of duodeno-jejunal flexure. Hematemesis is the cardinal sign (vomiting of blood or coffee ground-like material) though some children may present with malena (black, tarry stools). Hematochezia (passage of bright red blood in stools) is usually a feature of lower gastrointestinal bleeding (GIB), but some infants with UGIB can sometimes present with passage of bright red blood from the rectum because of rapid gastrointestinal transit in a briskly bleeding child.

Etiology

The etiology of UGIB varies with age but considerable overlap exists between the different age groups [1]. Rate and extent of bleeding is chiefly governed by the etiology of UGIB. In India portal hypertension has been reported to be the commonest cause (95 %) in contrast to the developed countries where non-variceal causes like peptic ulcer, esophagitis bleeding are common (66 %). Extra-hepatic portal vein obstruction (EHPVO) was the most common cause of variceal bleed in Indian children [2, 3] (Table 1). While evaluating a child with upper GI bleed it is important to keep the specific etiologies at different ages in mind (Table 2).
Table 1

Causes of UGIB among Indian children presenting to a hospital (adapted from reference 2 & 3)

Cause

Yachha et al. 1996

Mittal et al. 1994

Varices

95 %

39.4 %

Esophagitis

-

23.7 %

Gastritis

1.3 %

7.2 %

Gastric ulcer

-

1.2 %

Duodenal ulcer

-

0.42 %

Esophageal ulcer

-

0.42 %

Henoch Schönlein purpura

1.3 %

-

ITP

1.3 %

-

Gastroduodenal artery aneurysm

1.3 %

-

Unknown

-

27.5 %

Table 2

Age wise distribution of etiology of UGIB in children (Adapted from reference 1)

Age group

Well appearing

Ill appearing

Neonates

Swallowed maternal blood

Hemorrhagic gastritis

Hemorrhagic disease of newborn

Necrotizing enterocolitis

Drugs- heparin, indomethacin

Gastric stress ulcers

Thrombocytopenia, platelet dusfunction

Disseminated intravascular coagulation

Infants

Reflux esophagitis

Hemorrhagic gastritis

Reactive gastritis

Gastric stress ulcers

Arteriovenous malformation

 

Children

Mallory-Weiss tear

Esophageal varices ( liver disease)

Reflux esophagitis

Hemorrhagic gastritis

Reactive gastritis

Stress ulcers

Neonates

UGIB is rare in the first mo of life, but if seen must be distinguished from swallowed maternal blood. Hemorrhagic disease of the newborn due to vitamin K deficiency should be considered in neonates not given vitamin K prophylaxis at birth. Maternal idiopathic thrombocytopenia and maternal NSAID use can also cause bleeding in the newborn. Other causes include stress gastritis or ulcers, vascular anomalies, coagulopathy caused by infection, liver failure, or a congenital coagulation factor deficiency.

Infants and Toddlers

Stress ulcers and gastritis in a sick infant or toddler, peptic ulcers, variceal bleeding in children with portal hypertension and vascular malformations can cause major bleeds in this age group. Reflux esophagitis, esophageal or gastrointestinal foreign body, communicating duplication cysts, NSAIDS and corrosive injury can cause non life threatening bleeds. History of a choking episode, even if it was transient or occurred days or even weeks before the bleeding episode, is an important clue for a foreign body.

Older Children and Adolescents

Causes of UGIB in older children and adolescents are similar to those seen in adults. Varices, peptic ulcers, Dieulafoy’s lesions and vascular malformations can cause major bleeds. Reflux esophagitis, Drug induced gastritis, Mallory Weiss tears, communicating duplication cysts, stromal tumors, lymphomas (rarely), Crohn’s disease and portal hypertensive gastropathy can cause minor bleeds. Haemobilia or bleeding from the pancreas (splenic artery aneurysm) should be considered in the setting of trauma, liver biopsy, and acute/chronic pancreatitis. Henoch-Schönlein purpura and vasculitis such as polyarteritis nodosa are rarer causes of UGIB in this age group.

Key Issues

Initial Assessment and Stabilization

As for any other emergency the first priority should be to assess the airway, breathing and circulation of a patient presenting with UGIB. The most important aspect of the evaluation is to determine the degree and rapidity of blood loss. Orthostatic changes in blood pressure (more than 10 mm of Hg) suggest a moderate bleed (15–20 % of blood loss) and warrant a more aggressive approach to management. Presence of signs of shock (tachycardia, prolonged capillary refill time, cold-clammy skin, supine hypotension) indicates severe bleed of more than 25–30 % of blood loss and a need for immediate volume expansion and stabilization before proceeding to a diagnostic algorithm (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs12098-013-0987-x/MediaObjects/12098_2013_987_Fig1_HTML.gif
Fig. 1

Algorithmic approach to management of a patient with significant acute upper GI bleed

Once the patient is stable the following issues should be addressed before more elaborate diagnostic workup:
  1. 1.

    Whether actual blood or ingested substances

    The first consideration in evaluating children who have new-onset “bleeding” is to determine whether the material passed is actual blood. A number of substances may simulate bright red blood (food coloring, colored gelatin or children’s drinks, red candy, beets, tomato skins, rifampin, phenytoin, antibiotic syrups) or malena (bismuth or iron preparations, charcoal, spinach, blueberries, grapes, licorice). For detecting blood in vomitus or nasogastric aspirate, the Gastroccult® test is more accurate [4]. The test uses Gastroccult slides, sealed in special wrapper, and stored at room temperature (15–30 °C). One drop of gastric sample is placed to the occult blood test area of the slide and two drops of Gastroccult Developer is applied directly over the sample. The test is read within 60 s. The development of any trace of blue color in the occult blood test area is regarded as a positive for occult blood. Simultaneously, functionality of test should be tested. For this, one drop of Gastroccult Developer is added between the positive and negative Performance Monitor areas and results interpreted within 10 s. If the slide and developer are functional, blue color will appear in the positive area, and the negative area will remain colorless.

     
  2. 2.

    In a neonate- whether it is patient’s own blood or swallowed blood

    This is often applicable for neonates. The Apt-Downey Test is used to differentiate between swallowed maternal blood and patient’s blood.

     
  3. 3.

    Is there a pulmonary, oral or ENT source of bleed?

    Bleeding from these sources may mimic GI bleed especially when the blood is swallowed. It may be seen in conjunction with epistaxis, sore throat or may follow dental procedures or tonsillectomy. Hence these areas must be explored to rule out in cases of doubt.

     
  4. 4.

    Level of bleeding

    Vomiting of bright red or coffee ground vomitus is the classic presentation of upper GI bleeding. Bloody diarrhea and bright red blood mixed or coating normal stool are the classic presentations of lower GI bleeding. However, hematochezia, malena, or occult GI blood loss could represent upper or lower GI bleeding. In case of acute-onset hematochezia or malena, the level of bleeding can be confirmed by passage of a nasogastric tube. Presence of blood in the stomach and clearing of nasogastric aspirate with lavage are diagnostic of UGIB.

     

Diagnostic Evaluation

Focused History

History is directed at finding the likely source and cause of bleed. It should include the following:
  • Recent or recurrent epistaxis: it raises the possibility of a nasopharyngeal source of bleeding.

  • Recent onset of jaundice and/or change in stool color: may suggest underlying liver disease.

  • History of easy bruising or bleeding: may suggests a disorder of coagulation, platelet dysfunction, or thrombocytopenia.

  • History of any co-morbid illnesses: Personal or family history of liver, kidney or heart disease, or coagulation disorders.

  • Epigastric pain, food pain relationship, may be an indicator of gastritis, esophagitis or ulcer disease.

  • Details of recent medications ingested as well as an ‘over the counter’ or ‘alternative’ drug history is important to assess potential contributions from medications that may induce ulceration (such as NSAIDs and corticosteroids). Even short-term use of ibuprofen can cause gastric ulcers and hematemesis.

  • History of previous bleeding episodes can point to EHPVO, vascular malformations and duplication cysts.

Physical Examination

Focused physical examination, besides assessing the severity of bleed, should include a good inspection of skin (for petechiae and echymoses, and vascular malformations) and mucous membranes of the nose and throat (for bleeding), and palpation of abdomen for localized tenderness and organomegaly.
  • The presence of hepatosplenomegaly and prominent abdominal vessels and a protruding abdomen may indicate portal hypertension and bleeding from esophageal varices,

  • Epigastric tenderness might suggest acute gastritis or peptic ulcer disease.

  • Vascular malformations like hemangiomas and telangiectases should also be noted.

  • Clinical evidences of bleeding diathesis like petechiae, echymoses, purpura etc. should also be noted.

  • Stigmata of chronic liver disease

Investigations

In an emergency setting only a few laboratory tests are essential in the beginning to evaluate UGIB. These include complete blood count, prothrombin time (PT) and partial thromboplastin time (PTT), liver function tests and blood grouping and cross matching if there is significant bleed. Testing for H. pylori (urea breath test, stool antigen and serological tests) is indicated in a patient with a probability of peptic ulcer. Additional laboratory evaluation depends on the result of the initial evaluation, the patient’s response to treatment, and clinical suspicion of a particular diagnosis.

Radiographic Studies

Abdominal ultrasound is useful in patients where extra hepatic portal hypertension, portal hypertension due to liver disease, hemobilia, splenic artery aneurysm or large vascular anomalies are suspected. Doppler blood flow can identify evidence of cirrhosis and portal blood flow dynamics.

Endoscopy

Upper gastrointestinal endoscopy is the gold standard for diagnosis and treatment of UGIB. It is the procedure of choice for all patients with UGIB. In the hands of skilled endoscopist, this procedure now can diagnose the etiology in 85– 90 % of cases. It is indicated to identify the site of the bleeding, to diagnose the specific cause of the bleeding, and to initiate therapeutic interventions when indicated. In case the endoscopic appearance suggests esophagitis, gastritis or duodenitis a biopsy should be obtained; in suspected peptic ulcer disease antral biopsies for H. pylori work up (histological examination, rapid urease test, and culture) should be taken. Though there is no definite time frame given, in all cases of major upper GI bleed, an early endoscopy (within first 24 h) is recommended by most of the reviews [5]. Endoscopy is contraindicated in hemodynamically unstable patients; it should be considered only after stabilization because early endoscopy (<6 h after presentation) in such patients has not improved outcomes.

CT Angiography

CT angiography may help to delineate vascular malformations beyond the duodenum, in areas not accessed by routine upper GI endoscope.

Nuclear Scintigraphy

In patients with persistent bleeding in whom endoscopy fails to identify a bleeding site, radioisotope-tagged red blood cell scans using technetium 99 m-sulfur colloid may be capable of detecting the site of bleeding. This modality is useful only if the rate of bleeding exceeds 0.1 ml/min. However, this modality has significant false localization and false-negative rates [1].

Angiography

Celiac/ Superior mesenteric artery angiography is used selectively in children with non variceal massive bleeding, e.g., from a peptic ulcer, that obscures endoscopic evaluation and therapy. It is also very useful in Hemobilia, splenic artery aneurysms and some types of vascular malformations. Bleeding must be more than 0.5 ml/min to be detected by angiography [6]. Angiography is not only helpful in making the diagnosis but also facilitates embolic coil/fibric/glue occlusion of the arterial branches supplying the lesion in case of vascular malformation.

Some newer investigations are available in a few centres for the evaluation of upper GI bleed from obscure areas beyond the duodenum (Capsule Endoscopy/ Double Balloon Enteroscopy) and in the liver or pancreas (Endosonography) in older children.

Management

The initial steps in management of severe acute GI Bleed include assessment, resuscitation, re-evaluation, identification of the cause and source of bleeding, and commencing appropriate treatment (Fig. 1). This requires a multidisciplinary approach. Early consultation from pediatric gastroenterologist is recommended for patients who have active ongoing bleed. The Interventional Radiologist and the Pediatric Surgeon must be kept in the picture in case of the failure/non-feasibility of endoscopic therapy, massive bleeding, recurrent bleeding, bleeding associated with significant abdominal pain. Care in the Pediatric Intensive Care setting can be life saving in the early part of the management before specific therapy can be instituted.

Resuscitation and Stabilization

Children presenting with symptoms and signs of upper GI bleeding require prompt recognition during triage so that timely resuscitation can be initiated and hemodynamic stability restored.

Airway

Treatment should begin with airway protection. Intubation is indicated in obtunded patients or in those with uncontrolled massive hemetemesis to prevent aspiration and to facilitate upper endoscopy if it is necessary for bleeding control.

Breathing

Supplemental oxygen should be provided to all patients. The adequacy of the breathing efforts must be assessed and supported if needed.

Circulation

Patients are assessed for hypovolemia and shock to determine requirements for fluid infusion and transfusion of packed erythrocytes. Large bore venous access should be obtained to restore blood volume. Preferably 2 cannulae should be in place: one for fluid/blood resuscitation and the second for sampling and drugs. Peripheral venous access may be difficult in a child with hypovolemic shock due to peripheral vasoconstriction, therefore intraosseus access should be considered for large volume resuscitation. Subsequently central venous access may be required. Fluid resuscitation should begin with crystalloids (20 ml/ kg) while the blood is being arranged. The rate of blood transfusion is determined by the severity of the hypovolemia, continuing active bleeding, and presence of co-morbidities. Blood transfusion is not needed in a hemodynamically stable patient that has a hematocrit above 24 % at presentation. Over transfusion (volume overexpansion) should be avoided especially in variceal bleed. Hematocrit should not exceed 30 %. This can increase the portal venous pressure and aggravate the UGIB.

Reassessment and Monitoring

Vitals should be monitored every 10 min –15 min till the child is stabilized and then hourly for 24 h after stoppage of bleeding or stabilization.

Nasogastric Aspiration

Nasogastric aspiration and saline lavage are indicated in all patients with UGIB to confirm the presence of intragastric blood; to determine the rate of gross bleeding; to check for ongoing or recurrent bleeding; to clear gastric field for endoscopic visualization; to prevent aspiration of gastric contents; to prevent hepatic encephalopathy in patients of cirrhosis.

Lavage with iced or cooled solutions has not shown any recognized advantage over room temperature solutions. Some observers believe that solutions at 32 °C may interfere with local coagulation mechanisms [7].

Correction of Coagulopathies

Parenteral vitamin K should be administered empirically even when results of coagulogram are pending (infants, 1–2 mg/dose; children, 5–10 mg/ dose). Coagulopathy with an international normalized ratio (INR) higher than 1.5 or abnormal partial thromboplastin time (PTT) should be corrected with fresh frozen plasma (10 ml/kg initially); cryoprecipitates may be tried in the face of severe coagulopathy especially if volume of fluid has to be restricted; Factor VIIa has little additional advantage, even in chronic liver disease, and is not routinely recommended; Platelets transfusion is also not recommended unless there is active bleeding with low platelet counts. All these products should be included in the calculated resuscitation fluids.

Pharmacotherapy

Variceal Bleed

Pharmacological therapy has the advantages of being generally applicable and capable of being initiated as soon as a diagnosis of variceal hemorrhage is suspected. A meta-analysis comparing emergency sclerotherapy and pharmacological treatment shows a similar efficacy with fewer side effects with the latter thereby suggesting that pharmacological therapy should be considered first-line treatment of variceal bleeding [8].

Octreotide is a somatostatin analog that decreases splanchnic blood flow and has fewer hemodynamic adverse effects than vasopressin. Pediatric studies have shown that it controls UBIG in up to 70 % of children [9]. It is the drug of choice for variceal bleeds and initiated as a bolus injection of 1 mcg/kg (up to a maximum of 50 mcg) followed by continuous infusion of 1 mcg/kg per h, which may be increased hourly by 1 mcg/kg per h up to 4 mcg/ kg per h [1]. Infusion should be continued for at least 24–48 h after the bleeding has stopped to prevent recurrence. The major adverse effect of octreotide is hyperglycemia.

Vasopressin and Terlipressin

Vasopressin use has largely been replaced by octreotide and now by Terlipressin. The usual dose is 0.002 to 0.005 units/kg per min for 12 h, and then tapered over 24 to 48 h (maximum, 0.2 units/min) [1]. Its use is limited by the side effects of vasoconstriction. Nitroglycerin has also been used to decrease portal pressure and, when used in conjunction with vasopressin, may ameliorate some of its untoward effects. Vasopressin has been replaced by its longer acting and safer analogue Terlipressin which has been found to be as effective as Octreotide in adults. Experience with Terlipressin in children is limited though it is expected to be equally effective. An advantage is intermittent 4–6 hourly dosing.

Somatostatin is also used for control of active bleed, in a dose of 250 mcg/kg IV bolus followed by 250 mcg/kg/h continuous infusion. In case of response, the infusion can be maintained for 2 to 5 d, while frequently monitoring for hyperglycemia [1]. Side effects include abdominal discomfort, flushing, nausea, bradycardia, steatorrhoea and dyspepsia.

Prokinetic Agents

Erythromycin has been used as a prokinetic agent to clear the stomach of blood prior to emergent endoscopy. Metoclopramide has also been used to act as a prokinetic for similar reasons besides acting as a ‘pharmacologic tamponade’ – it increases the lower esophageal sphincter tone.

Mucosal Bleeds

Mucosal bleeding is most common type of upper GI bleed in critically ill children. Therapy in these groups of patients is directed at neutralizing and/or preventing the release of acid. The various agents used include:
  • Proton Pump Inhibitors (PPIs) are more efficacious than H2 receptor antagonists. Pantoprazole is used for control of active bleed. Dosage in children are: for body weight <40 kg: 0.5 to 1 mg/kg per day IV once daily; >40 kg: 20 to 40 mg once daily (maximum, 40 mg/d). These can be started empirically as it is important to raise the gastroduodenal pH to maintain clot stability. High dose PPI infusion has been found to decrease the need for endoscopic therapy.

  • H2Receptor Antagonists are used for control of active bleed and prevention of rebleeds. E.g. Ranitidine can be used as either continuous or bolus infusion; in the former 1 mg/kg is given initially followed by infusion of 2 to 4 mg/kg per day while in the latter 3 to 5 mg/kg per day is divided into every 8 hourly infusions.

  • Treatment for H. pylori infection with a H2 blockers or PPIs plus any two antibiotics (mainly nitromidazoles-metronidazole/tinidazole, macrolides- clarithromycin, amoxicillin and beta-lactames) for 10–14 d is recommended in patients with peptic ulcer disease positive for H. pylori or with no identifiable cause. American College of Gastroenterology recommends four specific drug regimens that use above referred combination of three medications [10] or Bismuth- containing quadruple therapy Bismuth subsalicylate, metronidazole, tetracycline all in 4 daily doses, and ranitidine or PPI twice a day for 10–14 d. These combinations are expected to cure 70 % to 85 % of infections [10].

Endoscopic Techniques

Variceal Bleed

Upper gastrointestinal endoscopy should be performed as soon as possible after initial stabilization. Endoscopic therapy should be done if variceal source of hemorrhage is confirmed. A meta-analysis has shown that endoscopic variceal ligation (EVL) is superior to sclerotherapy in the initial control of bleeding. However EVL cannot be performed in infants and toddlers due to the large size of available devices; EST is the mainstay of therapy in this group. Combination of pharmacological and endoscopic therapy is the most rational approach in the treatment of acute variceal hemorrhage [11]. Endoscopic injection of ‘tissue glue’ is effective for controlling bleeding gastric varices. Argon Plasma Coagulation can be used for bleeding portal hypertensive gastropathy lesions, e.g., GAVE- Gastric Antral Vascular Ectasias.

Non Variceal Ulcer Bleeds

In case of peptic ulcers with stigmata indicating high risk of rebleed (spurting or oozing vessel in ulcer base/ adherent clot), any of the following endoscopic therapy may be given: Injection therapy with adrenaline and saline, mechanical hemostasis (Endoclip Devices) with or without adrenaline, or thermocoagulation with or without adrenaline.

Balloon Tamponade

In patients with variceal bleed who continue to bleed despite pharmacologic and endoscopic methods, a Sengstaken-Blakemore tube can be placed to stop hemorrhage by mechanically compressing esophageal and gastric varices. However, its use is associated with potentially lethal complications such as aspiration, migration, and necrosis/perforation of the esophagus with mortality rates as high as 20 % [12]. The tube should never be kept inflated for durations exceeding 12 h in children.

Surgical Treatment

Consultation with a pediatric surgeon and interventional radiologist is necessary for children with massive bleeding, ongoing significant blood loss, and failed endoscopic procedure. Approaches to manage refractory variceal hemorrhage include insertion of transjugular intrahepatic porto-systemic shunt stents (in sinusoidal and post sinusoidal portal hypertension), selective or non-selective surgically created portosystemic shunts, and nonshunt procedures aimed at interrupting and ligating varices directly (devascularization) [13]. Non variceal bleed can be tackled by transcatheter embolization by an intervention radiologist. If this is not technically feasible or expertise is unavailable, surgical ligation / resection can be resorted to.

Key Points

  • Upper GI bleeding is often a medical emergency.

  • Specific etiologies at different ages should be kept in mind while assessing pediatric patients.

  • Variceal bleed is the most common cause of significant upper GI bleeding in children.

  • Immediate stabilization should be given priority before proceeding to diagnostic algorithm.

  • Upper GI endoscopy is the gold standard for diagnosis and treatment of UGIB.

  • Therapy in patients with mucosal bleeds is directed at neutralization and/or prevention of the gastric acid release. Proton Pump Inhibitors (PPIs) are more efficacious than H2 receptor antagonists.

  • Octreotide and terlipressin are useful in control of significant UGIB especially due to variceal hemorrhage.

  • Refractory variceal hemorrhage or non-variceal hemorrhage requires multidisciplinary approach.

Conflict of Interest

None.

Role of Funding Source

None.

Copyright information

© Dr. K C Chaudhuri Foundation 2013