Upper GI Bleeding
Bleeding proximal to the ligament of Treitz. Commonest causes: peptic ulcer disease (35-50%), oesophageal varices (10-20%), Mallory-Weiss tear (5-10%). Incidence ~100 per 100,000/year. Overall mortality ~10%.
Key Facts
Commonest causes: peptic ulcer (35-50%), varices (10-20%), oesophagitis/erosive gastritis (10-15%), Mallory-Weiss tear (5-10%), malignancy (2-5%) NICE NG141: risk assessment with Glasgow-Blatchford score at presentation; OGD within 24 hours (within 12 hours if variceal suspected or haemodynamically unstable) Forrest classification (peptic ulcer bleeding): Ia (spurting) and Ib (oozing) = active bleeding → endoscopic therapy; IIa (visible vessel) → endoscopic therapy; IIb (adherent clot) → consider; IIc/III (flat spot/clean base) → no endoscopic therapy Post-endoscopy PPI: high-dose IV omeprazole (80 mg bolus + 8 mg/hr × 72 hours) for Forrest Ia-IIb — reduces rebleeding Variceal UGIB: terlipressin 2 mg IV stat + antibiotics + urgent OGD band ligation + consider early TIPSS if high-risk (Child C 10-13 or Child B with active bleed) Restrictive transfusion: target Hb 70-80 g/L — reduces rebleeding and mortality (Villanueva et al., NEJM 2013)
Overview
Key Facts
Upper GI bleeding (UGIB) accounts for ~80% of acute GI bleeding and is a common medical emergency. Advances in endoscopic therapy and PPI use have improved outcomes, but mortality remains ~10% due to an aging comorbid population.
Epidemiology
Incidence ~100-150 per 100,000 per year in UK. ~70,000 hospital admissions/year. Mortality ~10% (higher in inpatient bleeding ~30%). Peptic ulcer disease is the commonest cause overall.
Aetiology
- Peptic ulcer disease: 35-50% (DU > GU)
- Oesophageal/gastric varices: 10-20%
- Erosive oesophagitis/gastritis: 10-15%
- Mallory-Weiss tear: 5-10% (post-vomiting/retching mucosal tear at GOJ)
- Upper GI malignancy: 2-5%
- Dieulafoy lesion: 1-2% (aberrant submucosal artery with small overlying mucosal defect)
- Angiodysplasia: 1-2%
- Aorto-enteric fistula: rare but catastrophic (history of aortic graft)
Pathophysiology
Peptic ulcer bleeding: ulcer erodes into a blood vessel (usually posterior duodenal ulcer eroding into gastroduodenal artery). Variceal bleeding: see oesophageal varices entry. Mallory-Weiss: longitudinal mucosal tear at GOJ from forceful vomiting/retching; usually self-limiting.
Clinical Presentation
Presentation
- Haematemesis: fresh red blood (active/severe) or coffee-ground (slower/altered)
- Melaena: black, tarry, offensive stools
- Haematochezia: fresh PR blood (if massive UGIB — brisk transit)
- Syncope/pre-syncope (from blood loss)
- Signs of shock: tachycardia, hypotension, pallor, cold peripheries
Red Flags
- Haemodynamic instability (HR >100, SBP <100)
- Fresh ongoing haematemesis
- Known/suspected varices (liver disease stigmata)
- Anticoagulated patient
- Aortic graft history (aorto-enteric fistula — often herald bleed then massive)
- Witnessed collapse with haematemesis
Differential Diagnosis
| Cause | Key Features | Endoscopic Finding |
|---|---|---|
| Peptic ulcer | Epigastric pain, H. pylori, NSAIDs | Ulcer ± stigmata of recent haemorrhage |
| Oesophageal varices | Liver disease, haematemesis | Variceal columns with bleeding point |
| Mallory-Weiss tear | Post-vomiting/retching, usually self-limiting | Longitudinal mucosal tear at GOJ |
| Erosive gastritis | NSAIDs, alcohol, stress (ICU) | Diffuse mucosal erosions |
| Malignancy | Weight loss, dysphagia, anaemia | Mass/ulcerated lesion |
| Dieulafoy lesion | Massive bleed, small mucosal defect | Visible vessel without surrounding ulcer |
Diagnosis / Investigation
Bedside
- A-E assessment: pulse, BP, SpO₂, RR, GCS
- DRE: melaena confirmation
- Glasgow-Blatchford Score: at presentation — determines urgency and disposition
Bloods
- FBC: Hb (may lag behind blood loss initially)
- U&Es: urea raised disproportionately in UGIB
- LFTs: liver disease assessment
- INR/coagulation: anticoagulant therapy, liver disease
- Group and crossmatch: 2-6 units
- Lactate: tissue perfusion
Endoscopy
- OGD: within 24 hours (within 12 hours if variceal suspected, haemodynamically unstable after resuscitation, or high-risk features)
- Forrest classification of ulcer bleeding guides management
- Endoscopic therapy: adrenaline injection (1:10,000 — for initial haemostasis) PLUS thermal or mechanical (clips) for definitive haemostasis; band ligation for varices
Imaging
- CT angiography: if massive bleeding and endoscopy unable to localise/control
- Mesenteric angiography: diagnostic and therapeutic (embolisation)
- CT thorax/abdomen: if aorto-enteric fistula suspected
Management
Non-pharmacological
- Resuscitation: 2 large-bore cannulae, IV crystalloid, crossmatch
- Restrictive transfusion: target Hb 70-80 g/L (Villanueva et al. NEJM 2013 — improves survival vs liberal transfusion, especially in variceal bleeding)
- Correct coagulopathy: vitamin K 5-10 mg IV for warfarin; idarucizumab for dabigatran; prothrombin complex concentrate for life-threatening bleeding on VKAs
- Keep NBM until endoscopy
Pharmacological
- Non-variceal UGIB:
- Pre-endoscopy PPI: IV omeprazole 80 mg bolus (does not reduce mortality but may reduce high-risk endoscopic stigmata)
- Post-endoscopy (high-risk — Forrest Ia-IIb): IV omeprazole 80 mg bolus then 8 mg/hr × 72 hours, then switch to oral PPI
- H. pylori test and treat: all peptic ulcer patients
- Review/stop NSAIDs: if NSAID-related; co-prescribe PPI if NSAID essential
- Variceal UGIB:
- Terlipressin 2 mg IV stat then 1-2 mg QDS × up to 5 days
- Prophylactic antibiotics: ceftriaxone 1 g IV OD × 5-7 days
- OGD band ligation: within 12 hours
- Early TIPSS: within 72 hours for high-risk (Child C 10-13 or Child B with active bleeding at OGD)
- Balloon tamponade (Sengstaken-Blakemore): temporary bridge if endoscopy unavailable or failed (max 24 hours)
Surgical/Interventional
- Rebleeding after endoscopic therapy: repeat OGD first; if fails → interventional radiology (embolisation) or surgery
- Surgery: under-running of bleeding vessel (duodenal ulcer → gastroduodenal artery), partial gastrectomy (rarely needed now)
- TIPSS: for variceal bleeding (see above)
Referral Criteria
- GBS 0: consider outpatient OGD (safe early discharge)
- GBS ≥1: inpatient management
- Variceal bleeding: urgent GI/hepatology
- Failed endoscopic haemostasis: IR + surgical standby
Prognosis
Overall UGIB mortality ~10%. Peptic ulcer bleeding rebleeding rate after endoscopic therapy ~5-10%. Variceal bleeding mortality 15-20% per episode. Mallory-Weiss tears: self-limiting in >90%. GBS 0 patients have <1% risk of requiring intervention. Inpatient UGIB has worse prognosis (~30% mortality) due to comorbidities. Restrictive transfusion strategy reduces mortality compared to liberal strategy.
Other Relevant Information
Forrest Classification of Peptic Ulcer Bleeding
| Class | Description | Rebleeding Risk | Endoscopic Therapy |
|---|---|---|---|
| Ia | Spurting haemorrhage | 55% | Yes |
| Ib | Oozing haemorrhage | 55% | Yes |
| IIa | Non-bleeding visible vessel | 43% | Yes |
| IIb | Adherent clot | 22% | Consider |
| IIc | Flat pigmented spot | 10% | No |
| III | Clean ulcer base | 5% | No |
Key Guidelines and Trials
| Guideline/Trial | Key Recommendation |
|---|---|
| NICE NG141 | GBS for risk assessment; OGD within 24h; restrictive transfusion |
| Villanueva et al. (2013) | Restrictive (Hb 70) vs liberal (Hb 90) — improved survival |
| BSG UGIB Guideline | Endoscopic dual therapy for high-risk ulcers; PPI infusion post-endoscopy |