Acute Upper GI Bleeding
Acute upper GI bleeding presents with haematemesis and/or melaena. Risk stratification with the Glasgow-Blatchford score guides management, with endoscopy within 24 hours and specific treatment for variceal and non-variceal sources.
Key Facts
~50,000-70,000 hospital admissions for upper GI bleeding per year in the UK; overall mortality ~6-8% Glasgow-Blatchford Score (GBS): GBS 0 = very low risk, can be considered for outpatient management; GBS ≥6 = high risk, requires endoscopy and admission Peptic ulcer disease is the most common cause (~35-50%); variceal bleeding accounts for ~10% but has the highest mortality (~15-20%) Endoscopy within 24 hours for all upper GI bleeds; within 2 hours for suspected variceal or haemodynamically unstable patients Rockall score: Post-endoscopy scoring system to predict rebleeding and mortality PPI therapy: Omeprazole 80mg IV bolus then 8mg/hr infusion for high-risk peptic ulcer bleeding post-endoscopy Variceal bleeding: IV terlipressin 2mg QDS + prophylactic antibiotics (co-amoxiclav) + band ligation at endoscopy Restrictive transfusion: Target Hb 70-80g/L (Villanueva trial — restrictive strategy improves survival in GI bleeding)
Overview
Key Facts
Acute upper GI bleeding is a common medical emergency. Risk stratification, appropriate resuscitation, and timely endoscopy are the cornerstones of management. The Blatchford score determines whether patients need admission, while the Rockall score predicts rebleeding and mortality.
Epidemiology
Upper GI bleeding accounts for approximately 50,000-70,000 admissions per year in the UK. Overall mortality is approximately 6-8% but varies significantly: peptic ulcer ~5%, variceal bleeding ~15-20%. Incidence increases with age. Approximately 80% of non-variceal bleeds stop spontaneously.
Aetiology
- Peptic ulcer disease (~35-50%): Gastric and duodenal ulcers; H. pylori, NSAIDs
- Oesophageal varices (~10%): Portal hypertension (liver cirrhosis)
- Mallory-Weiss tear (~10%): Mucosal tear at GOJ from retching/vomiting
- Oesophagitis/gastritis (~15%): Reflux, alcohol, NSAIDs
- Upper GI malignancy (~5%): Gastric, oesophageal
- Dieulafoy lesion, AVM, aorto-enteric fistula: Rare but important
Pathophysiology
Upper GI bleeding originates proximal to the ligament of Treitz. Peptic ulcer bleeding results from erosion into submucosal or deeper vessels by acid-pepsin injury. Variceal bleeding occurs from rupture of dilated portosystemic collateral veins due to portal hypertension (portal pressure >12 mmHg). The haemodynamic response includes tachycardia, hypotension, and compensatory vasoconstriction.
Clinical Presentation
Presenting Features
- Haematemesis: Fresh blood (bright red) or coffee-ground vomiting
- Melaena: Black, tarry, offensive stools (>50mL blood in GI tract)
- Haematochezia: Fresh rectal bleeding (massive upper GI bleed with rapid transit)
- Dizziness, presyncope, or syncope
- Epigastric pain (peptic ulcer)
Red Flags
- Haemodynamic instability (SBP <100, HR >100) — significant haemorrhage
- Fresh haematemesis — active bleeding
- Suspected variceal bleed (known liver disease, stigmata of chronic liver disease)
- Anticoagulated patient — higher bleeding risk
- Coffee-ground vomiting with tachycardia — may underestimate blood loss
- Aorto-enteric fistula ('herald bleed' followed by massive haemorrhage) — consider if previous aortic graft
Differential Diagnosis
| Diagnosis | Key Features | Investigation |
|---|---|---|
| Peptic ulcer | Epigastric pain, NSAID/H. pylori history | OGD — Forrest classification |
| Variceal bleed | Liver disease, massive haematemesis | OGD — band ligation |
| Mallory-Weiss tear | Vomiting/retching then haematemesis | OGD |
| Oesophagitis/gastritis | Reflux symptoms, alcohol, NSAIDs | OGD |
| Gastric cancer | Weight loss, early satiety, elderly | OGD + biopsy |
| Aorto-enteric fistula | Previous aortic surgery, massive bleed | CT angiography |
Diagnosis / Investigation
Bedside
- ABCDE assessment: Haemodynamic stability
- Glasgow-Blatchford Score: Determines need for admission and intervention
- Blood glucose: Baseline
- DRE: Melaena confirmation
Bloods
- FBC: Hb (may be normal initially in acute bleed — haemoconcentration), platelets
- U&Es: Raised urea disproportionate to creatinine (absorbed blood proteins)
- LFTs: If liver disease/variceal bleed suspected
- Coagulation (INR): Anticoagulated patients, liver disease
- Group and crossmatch: 2-6 units depending on severity
- Lactate: Tissue perfusion marker
Imaging
- OGD (oesophagogastroduodenoscopy): Within 24h (within 2h if variceal or haemodynamically unstable)
- CT angiography: If endoscopy fails to identify source or is not feasible
- Mesenteric angiography: Therapeutic embolisation if endoscopy fails
Special Tests
- H. pylori testing: Urease test at endoscopy, stool antigen, or urea breath test
- Forrest classification: Endoscopic classification of peptic ulcer bleeding risk
Management
Non-pharmacological
- Resuscitation: ABCDE, large-bore IV access (×2), fluid resuscitation
- Restrictive transfusion: Target Hb 70-80g/L (Villanueva trial — liberal transfusion increases rebleeding and mortality)
- Correct coagulopathy: FFP if INR >1.5; platelets if <50 × 10⁹/L; vitamin K if warfarin
Pharmacological
Non-variceal bleeding:
- PPI: IV omeprazole 80mg bolus then 8mg/hr infusion for 72h if high-risk ulcer at endoscopy
- H. pylori eradication: If positive — triple therapy (e.g., amoxicillin 1g BD + clarithromycin 500mg BD + PPI for 7 days)
- Stop NSAIDs: Permanently if possible; lowest dose + PPI if essential
Variceal bleeding:
- IV terlipressin 2mg QDS (reduces portal pressure) — continue for up to 5 days
- Prophylactic antibiotics: Co-amoxiclav 1.2g IV TDS (or ciprofloxacin if penicillin allergic) for 5 days — reduces mortality
- Endoscopic band ligation: Within 2 hours; variceal banding for oesophageal varices
- Sengstaken-Blakemore tube: Temporary tamponade if bleeding uncontrolled pending endoscopy/TIPSS
Anticoagulated patients:
- Warfarin: Vitamin K + PCC (Beriplex) — target INR <1.5
- DOAC: Discuss with haematology; may need specific reversal agents
Surgical/Interventional
- Endoscopic therapy: Clip application, adrenaline injection, thermal coagulation, band ligation
- Interventional radiology: Embolisation for refractory non-variceal bleeding
- TIPSS (Transjugular Intrahepatic Portosystemic Shunt): For refractory or recurrent variceal bleeding
- Surgery: Rare — oversewing of bleeding ulcer if endoscopy and IR fail
Referral Criteria
- GBS ≥6 — admission, endoscopy within 24h
- GBS 0 — consider outpatient endoscopy and discharge with safety net
- Suspected variceal bleed — urgent endoscopy within 2h, hepatology involvement
- Rebleeding after endoscopic therapy — repeat endoscopy, consider IR or surgery
Prognosis
- Overall mortality: ~6-8% for all upper GI bleeding
- Peptic ulcer: ~5% mortality; ~15-20% rebleed rate after endoscopic therapy
- Variceal bleeding: ~15-20% mortality per episode; ~60% rebleed within 1 year without secondary prevention
- GBS 0: Very low risk (<1%) of adverse outcome — safe for outpatient management
- Restrictive transfusion: Reduces 45-day mortality (Villanueva RCT)
- H. pylori eradication: Reduces ulcer recurrence from ~60% to <5%
Other Relevant Information
Glasgow-Blatchford Score (GBS)
| Parameter | Score |
|---|---|
| Hb (men): 120-129 → 1; 100-119 → 3; <100 → 6 | |
| Hb (women): 100-119 → 1; <100 → 6 | |
| Urea: 6.5-7.9 → 2; 8-9.9 → 3; 10-25 → 4; >25 → 6 | |
| SBP: 100-109 → 1; 90-99 → 2; <90 → 3 | |
| HR ≥100 → 1 | |
| Melaena → 1 | |
| Syncope → 2 | |
| Liver disease → 2 | |
| Heart failure → 2 |
Forrest Classification of Peptic Ulcer Bleeding
| Class | Appearance | Rebleed Risk |
|---|---|---|
| Ia | Spurting haemorrhage | ~55% |
| Ib | Oozing haemorrhage | ~55% |
| IIa | Visible vessel (non-bleeding) | ~43% |
| IIb | Adherent clot | ~22% |
| IIc | Flat pigmented spot | ~10% |
| III | Clean base ulcer | ~5% |