Paediatric upper gastrointestinal (UGI) haemorrhage is a low-frequency but high-acuity presentation that sits squarely within the FRCEM SBA paediatric revision curriculum. Its clinical complexity arises from the breadth of age-dependent aetiology, the risk of misclassification as a benign complaint, and the need to integrate APLS resuscitation principles with condition-specific pharmacology including octreotide and prophylactic antibiotics. Examiners use this topic to probe structured diagnostic reasoning, appropriate escalation, and evidence-based prescribing under pressure. Candidates who approach it systematically consistently outperform those relying on rote recall.
Key Points: Upper GI Bleeding in Children
- Aetiology is strongly age-dependent: swallowed maternal blood and haemorrhagic disease of the newborn dominate in neonates; oesophageal varices and peptic ulcer disease are the high-stakes diagnoses in older children.
- Distinguish true UGI haemorrhage from swallowed blood (maternal or epistaxis) using the Apt-Downey test in neonates and targeted history in older children before committing to an aggressive workup.
- Button battery ingestion is a UGI emergency: liquefactive necrosis can develop within two hours; immediate plain radiography and emergency endoscopy or retrieval are mandatory.
- Variceal haemorrhage management requires octreotide infusion, prophylactic antibiotics (ceftriaxone), blood product resuscitation targeting haemoglobin 70-80 g/L, and urgent endoscopy within 12 hours.
- Proton pump inhibitors are first-line for non-variceal UGI bleeding; high-dose intravenous omeprazole reduces rebleeding risk in peptic ulcer disease.
- Early specialist involvement (paediatric gastroenterology, paediatric surgery, and critical care) is essential in any child with haemodynamic compromise or suspected varices.
Clinical Context and Epidemiology
UGI bleeding in children is uncommon, with population-based estimates suggesting an annual incidence of approximately 1 to 2 per 10,000 children. Despite lower overall mortality compared with adult UGI haemorrhage, outcomes deteriorate significantly when portal hypertension underlies the presentation: reported mortality from variceal bleeding episodes in children with portal hypertension approaches 2 to 5% per episode. This figure rises further in critically ill children and those with decompensated chronic liver disease. The Royal College of Paediatrics and Child Health acknowledges paediatric gastrointestinal emergencies as a core competency domain, reflecting the clinical and medicolegal stakes of delayed recognition.
From an ED triage perspective, the presenting complaint is often haematemesis or melaena, but the differential is wide and includes conditions that mimic GI bleeding without representing true mucosal haemorrhage. Swallowed maternal blood accounts for the majority of apparent haematemesis in neonates and is entirely self-limiting. Similarly, epistaxis with swallowed blood, haemoptysis, and ingested red food colourings can generate unnecessary resuscitation if the clinician does not apply a systematic approach from the outset.
Age-Specific Aetiology: The Examiner’s Favourite Framework
The single most reliable organising principle for FRCEM SBA paediatric revision questions on this topic is the age-stratified differential. Examiners consistently embed the critical diagnosis within an age-appropriate clinical vignette, and candidates who internalise this framework answer these questions more quickly and accurately.
Neonates (0 to 28 days)
Swallowed maternal blood is the most common cause of apparent haematemesis in neonates and is distinguished from neonatal haemorrhage using the Apt-Downey test: fetal haemoglobin (HbF) resists alkaline denaturation and remains pink, whereas maternal HbA turns yellow-brown. Haemorrhagic disease of the newborn (vitamin K deficiency bleeding) must be considered in any neonate not given prophylactic vitamin K at birth. Stress ulcers, oesophagitis, Mallory-Weiss tears related to delivery, and coagulopathy of prematurity complete the neonatal differential.
Infants (1 to 12 months)
Gastro-oesophageal reflux-related oesophagitis and gastritis are common at this age. However, the most dangerous diagnosis is oesophageal varices secondary to extrahepatic portal hypertension, most often caused by neonatal portal vein thrombosis. Cow’s milk protein allergy can cause mucosal inflammation with blood in vomitus, though this is more characteristically a lower GI presentation.
Toddlers and Pre-school Children (1 to 5 years)
This age group carries the highest risk from button battery ingestion, a true UGI emergency. Lodged batteries generate a hydroxide-mediated liquefactive necrosis within as little as two hours, with oesophageal perforation, tracheo-oesophageal fistula, and fatal haemorrhage reported in the literature. Immediate erect chest and abdominal radiography is mandatory; endoscopic or surgical retrieval cannot await routine scheduling. Portal hypertension from biliary atresia (including post-Kasai portoenterostomy failure) and Mallory-Weiss tears post-vomiting are also seen in this group.
School-age Children (5 to 12 years)
Helicobacter pylori-associated peptic ulcer disease becomes clinically relevant from school age onwards. Oesophageal varices remain important, and rarer causes including Dieulafoy lesions (a submucosal arteriole eroding through the mucosa without ulceration) and Henoch-Schonlein purpura with UGI mucosal involvement should enter the differential when the history does not fit more common diagnoses.
Adolescents (12 to 16 years)
The adolescent differential closely resembles adult UGI bleeding. Peptic ulcer disease, NSAID and aspirin gastropathy, and Mallory-Weiss tears (particularly in the context of alcohol use or eating disorders with recurrent vomiting) predominate. Chronic liver disease from conditions such as autoimmune hepatitis, Wilson’s disease, or cystic fibrosis-related liver disease must be excluded in any adolescent with varices.
ED Assessment: History, Examination, and Investigations
Targeted History
Key historical features that stratify risk and guide diagnosis include: known liver disease, biliary atresia, or prior Kasai portoenterostomy (high variceal risk); NSAID, aspirin, or corticosteroid use; recent forceful vomiting; foreign body ingestion history; inherited coagulopathy or haematological malignancy; prematurity and vitamin K status in neonates; and a family history of portal hypertension or coagulopathy.
Examination and Haemodynamic Assessment
Apply APLS principles throughout. Signs of haemodynamic compromise in children differ from adults: tachycardia is an early and sensitive indicator; hypotension is a late and pre-terminal sign. Compensated shock in a child with variceal haemorrhage can be catastrophically rapid in onset. Assess for stigmata of chronic liver disease (splenomegaly, caput medusae, jaundice, spider naevi, palmar erythema), which immediately raise variceal probability.
Investigations
First-line investigations include: full blood count (baseline haemoglobin and platelet count); coagulation screen including INR (deranged in liver disease and coagulopathy); urea and electrolytes (elevated urea-to-creatinine ratio supports UGI rather than lower GI blood loss); liver function tests; cross-match and group-and-save; and venous blood gas for rapid haemoglobin estimation and acid-base status. Plain radiography of chest and abdomen is essential if button battery ingestion is suspected. Nasogastric aspirate can be helpful to confirm UGI bleeding but a clear aspirate does not exclude it. Endoscopy remains the gold-standard diagnostic and therapeutic intervention and should be arranged urgently in haemodynamically significant or unresolved bleeding.
Management: Resuscitation and Condition-Specific Pharmacology
Resuscitation
Establish two large-bore intravenous cannulae. Initial fluid resuscitation follows Resuscitation Council UK and APLS guidance: 10 mL/kg isotonic crystalloid (0.9% sodium chloride) boluses, reassessing after each, aiming for haemodynamic stabilisation. Avoid excessive crystalloid in suspected variceal bleeding, where dilutional coagulopathy and elevated portal pressure worsen outcomes. Early activation of the major haemorrhage protocol and transfusion of packed red blood cells is appropriate in ongoing haemorrhage; target haemoglobin 70 to 80 g/L in variceal bleeding to avoid worsening portal pressures. Platelet and fresh frozen plasma transfusion should be guided by results and clinical trajectory.
Variceal Haemorrhage
When variceal bleeding is confirmed or strongly suspected, initiate octreotide (somatostatin analogue) as a continuous intravenous infusion. Octreotide reduces splanchnic blood flow and portal pressure, reducing active bleeding and facilitating endoscopy. Paediatric dosing is weight-based and should be confirmed against the British National Formulary for Children. Prophylactic antibiotics, typically intravenous ceftriaxone, are recommended to reduce infectious complications and improve survival in children with portal hypertension and variceal haemorrhage, consistent with adult evidence extrapolated to paediatric practice. Urgent upper GI endoscopy (ideally within 12 hours) allows variceal band ligation or sclerotherapy. Balloon tamponade (Sengstaken-Blakemore or Minnesota tube) is a temporising measure of last resort in massive uncontrolled haemorrhage when endoscopy cannot be immediately performed.
Non-Variceal Haemorrhage
For peptic ulcer disease and other non-variceal causes, high-dose intravenous proton pump inhibitor therapy (omeprazole or pantoprazole) is first-line. NICE guidance on acute upper gastrointestinal bleeding supports PPI use pre-endoscopy to reduce the proportion of patients with stigmata of recent haemorrhage at endoscopy, though its survival benefit is not proven. Post-endoscopic PPI infusion reduces rebleeding risk following haemostatic intervention for high-risk ulcers.
How the FRCEM SBA Tests This Topic
In FRCEM SBA paediatric revision questions on UGI bleeding, the examiner typically presents a short clinical vignette with age, presenting complaint (haematemesis or melaena), and one or two discriminating historical features. The SBA then asks the candidate to identify the most likely diagnosis, the most appropriate immediate investigation, or the correct initial pharmacological intervention.
Common question stems include: a jaundiced three-year-old with post-Kasai portoenterostomy history presenting with fresh haematemesis (expected answer: oesophageal varices, initiate octreotide); a six-week-old neonate with blood-stained vomit and a well appearance (expected answer: Apt-Downey test to exclude swallowed maternal blood); an 18-month-old with drooling and haematemesis after unsupervised access to a small coin-shaped object (expected answer: button battery on plain radiograph, emergency retrieval). Pitfalls include selecting a PPI as first-line in confirmed variceal haemorrhage (octreotide is the correct answer), or failing to recognise button battery urgency and selecting watchful waiting.
The RCEM curriculum maps paediatric gastrointestinal emergencies to core and higher-level competencies across the emergency medicine training pathway. Candidates preparing for the FRCEM SBA exam should ensure familiarity with the APLS-based resuscitation framework as well as the pharmacological specifics outlined above.
Revision Pearls: High-Yield Facts for FRCEM SBA Paediatric Revision
- The Apt-Downey test distinguishes fetal haemoglobin (alkali-resistant, remains pink) from maternal haemoglobin (turns brown) in neonatal haematemesis.
- Button battery ingestion causes liquefactive necrosis within two hours and demands emergency retrieval; never adopt a wait-and-see approach.
- Octreotide, not a proton pump inhibitor, is the pharmacological priority in suspected variceal haemorrhage.
- Prophylactic ceftriaxone reduces mortality in variceal haemorrhage with portal hypertension; this is extrapolated from adult trial data and is standard UK paediatric gastroenterology practice.
- A urea-to-creatinine ratio above 100:1 (in SI units, urea above 6.5 mmol/L with normal creatinine) supports a UGI rather than lower GI source.
- In children, hypotension is a pre-terminal sign of shock; tachycardia alone in the context of haematemesis mandates aggressive resuscitation without waiting for blood pressure to fall.
- Target haemoglobin of 70 to 80 g/L in variceal bleeding: over-transfusion raises portal pressure and worsens rebleeding risk.
- Any child with known biliary atresia, Kasai portoenterostomy, or chronic liver disease presenting with haematemesis should be presumed to have variceal bleeding until proven otherwise.
Common Pitfalls: Where Candidates Lose Marks
- Selecting PPI therapy as the priority intervention when the clinical scenario clearly points to variceal haemorrhage.
- Missing the button battery diagnosis by not obtaining a plain radiograph before reassuring a carer about coin ingestion.
- Failing to apply the Apt-Downey test concept in a neonatal haematemesis question, defaulting instead to endoscopy or resuscitation without establishing whether the blood is maternal or fetal in origin.
- Over-transfusing a child with suspected varices by aiming for a normal haemoglobin rather than a restrictive target.
- Not escalating to paediatric surgery and gastroenterology early enough in the clinical scenario; SBA distractors frequently include watchful waiting or observation as options.
- Confusing melaena (black, tarry, offensive stool indicating digested blood from a UGI source) with rectal bleeding (fresh red blood suggesting lower GI pathology), which changes the entire diagnostic and management pathway.
How EM Learning Centre Supports Your MRCEM Single Best Answer Revision
FRCEM SBA paediatric revision questions on topics such as upper GI bleeding require you to integrate age-specific clinical reasoning, pharmacological precision, and resuscitation frameworks under timed conditions. This is exactly the skill set that the MRCEM Single Best Answer revision course at EM Learning Centre is designed to build. The platform provides structured clinical lessons, high-quality SBA practice questions mapped directly to the RCEM curriculum, and detailed explanations that go beyond the answer to reinforce the underlying clinical science.
Whether you are consolidating a knowledge gap in paediatric emergencies, reviewing your approach to haemodynamic resuscitation, or working through a timed FRCEM SBA mock exam session, the EM Learning Centre offers the breadth and depth of content needed to perform consistently on exam day. Related lessons available on the platform, including those covering acute allergy, anaphylactoid reactions, angioedema, hereditary angioedema, and urticaria in the Emergency Department, illustrate the same principle: high-stakes low-frequency presentations require active, repeated retrieval practice rather than passive reading to achieve exam-ready retention.
If you are targeting the FRCEM SBA exam or working through your MRCEM SBA revision, systematic use of a question bank aligned to the current RCEM curriculum, combined with focused case-based learning, is the most evidence-consistent approach to preparation. Start with the topics that carry the highest examiner weighting and the greatest clinical risk, and build outwards from there.
References
- Royal College of Paediatrics and Child Health. Paediatric emergency medicine competency frameworks. rcpch.ac.uk
- National Institute for Health and Care Excellence. Acute upper gastrointestinal bleeding in over 16s: management. NICE guideline CG141. nice.org.uk
- Resuscitation Council UK. Advanced Paediatric Life Support guidance and resuscitation guidelines. resus.org.uk
- British National Formulary for Children. Octreotide and proton pump inhibitor dosing in paediatric patients. bnf.nice.org.uk
- Royal College of Emergency Medicine. Emergency medicine curriculum and FRCEM examination syllabus. rcem.ac.uk
- The BMJ. Clinical review: upper gastrointestinal bleeding in children. bmj.com
