---

Upper Gastrointestinal Bleeding - Complete Consultant Reference

---

1. HISTORY TAKING

A. Presenting Complaint - Characterize the Bleed

• Fresh red blood vs. coffee-ground material (digested blood indicates slower or stopped bleeding)
• Volume: estimate in cups/liters; number of episodes
• Presence of clots (suggests significant arterial hemorrhage)
• Association with retching or vomiting before the bleed (Mallory-Weiss tear)

• Black, tarry, offensive-smelling stool - confirms upper GI source in most cases
• Age <50 with melena more strongly suggests UGIB even without hematemesis
• Duration and number of episodes

• Bright red per rectum - typically lower GI, but ~14-15% of hematochezia cases originate from an upper GI source (usually massive UGIB)
• If hematochezia + hemodynamic instability - always consider upper GI source

• Iron or bismuth ingestion simulates melena (stool guaiac negative)
• Red-dyed liquid medications and beets simulate hematochezia (guaiac negative)
• Swallowed blood from epistaxis, dental procedures, or haemoptysis

B. Associated Symptoms

C. Haemodynamic Symptoms (Severity Assessment)

• Syncope or pre-syncope (significant volume loss >15-20%)
• Dizziness on standing (orthostatic symptoms)
• Palpitations, chest pain (cardiac ischaemia secondary to bleeding)
• Confusion, restlessness (cerebral hypoperfusion)
• Oliguria, anuria

D. Past Medical History

• Previous GI bleeding - prior episodes, their source, prior endoscopy findings and interventions (banding, injection sclerotherapy, clips)
• Peptic ulcer disease or dyspepsia
• Liver disease, hepatitis B/C, alcohol cirrhosis
• Inflammatory bowel disease
• Coagulopathy or haematological disorders
• Renal failure (uraemia causes platelet dysfunction; raised BUN:Cr ratio can indicate UGIB itself)
• Cardiovascular disease (relevant for transfusion threshold)
• Prior abdominal surgery or aortic aneurysm repair (aortoenteric fistula)
• Malignancy (known gastric, oesophageal, hepatic, or pancreatic cancer)
• Varices - known, previously banded, TIPS procedure

E. Drug History (Critical)

• NSAIDs and aspirin - most important risk factor for peptic ulcer bleeding; 57% of gastric ulcer bleeds and 53% of duodenal ulcer bleeds are associated with NSAID/aspirin use
• Anticoagulants - warfarin, DOACs (rivaroxaban, apixaban, dabigatran), heparin
• Antiplatelet agents - clopidogrel, ticagrelor, prasugrel (especially dual antiplatelet therapy)
• Corticosteroids - especially combined with NSAIDs (synergistic ulcer risk)
• SSRIs - inhibit platelet serotonin, impair haemostasis
• Bisphosphonates - pill-induced oesophagitis
• Tetracyclines, iron, potassium chloride - oesophageal/gastric erosions
• Thrombolytics - iatrogenic haemorrhage
• PPI/H2RA use - may partially mask symptoms

F. Social History

• Alcohol use - quantity and pattern (alcoholic cirrhosis, erosive gastritis, Mallory-Weiss)
• Smoking (impairs mucosal defense, NSAID use risk)
• Overseas travel (portal hypertension from schistosomiasis in endemic areas)
• Occupation and stress levels

G. Family History

• Hereditary haemorrhagic telangiectasia (HHT/Osler-Weber-Rendu)
• Peutz-Jeghers syndrome
• Familial polyposis
• Gastric malignancy

---

2. PHYSICAL EXAMINATION

A. Immediate Assessment - Haemodynamic Status

B. General Inspection

• Pallor (anaemia)
• Diaphoresis, cold clammy skin (shock)
• Agitation, confusion (cerebral hypoperfusion)
• Jaundice (liver disease)
• Cachexia (malignancy)
• Nutritional status

C. Skin and Peripheral Signs

• Exclude epistaxis, posterior nasal bleeding as source of swallowed blood producing coffee-ground emesis - this accounted for 2.2% of UGIB in the UCLA CURE database

D. Cardiovascular Examination

• Signs of heart failure (JVP, pulmonary oedema) - relevant for fluid resuscitation and Blatchford score
• Peripheral vascular disease (may coexist with aortic pathology)

E. Abdominal Examination

• Inspection: Scars (previous surgery, aortic repair), distension, visible peristalsis, caput medusae
• Palpation:
  • Epigastric tenderness (peptic ulcer, gastritis)
  • Hepatomegaly (cirrhosis, hepatic malignancy)
  • Splenomegaly (portal hypertension, haematological disorder)
  • Abdominal mass (gastric/pancreatic malignancy, lymphoma)
  • Pulsatile epigastric mass (AAA - risk of aortoenteric fistula)
  • Ascites (cirrhosis, malignancy)
• Percussion: Shifting dullness (ascites), liver dullness
• Auscultation: Bowel sounds (hyperactive in UGIB due to blood in intestine acting as cathartic)

F. Rectal Examination

• Examine stool colour: bright red (hematochezia), maroon, or black tarry (melena)
• Guaiac (FOB) testing of stool
• Assess for rectal masses, haemorrhoids

G. Nasogastric Aspirate (Bedside Procedure)

• Positive aspirate (bloody, maroon, or coffee-ground): confirms UGIB; high-risk lesions more likely
• Negative clear aspirate: does NOT exclude UGIB - intermittent bleeding, pyloric spasm, or oedema can prevent duodenal blood reflux; only positive in ~23% of patients with occult UGIB without hematemesis
• Bilious aspirate: suggests pylorus is patent; duodenal source is less likely if no blood
• Visual inspection is more reliable than guaiac testing of aspirate (traumatic insertion causes false positives)
• Note: No evidence that NG tube passage provokes variceal bleeding

---

3. INVESTIGATIONS

A. Immediate Bloods (Resuscitation Phase)

Digested blood is absorbed as urea, raising the BUN level. A BUN:Cr ratio ≥30 suggests an upper GI source even when the presentation is atypical. - Frameworks for Internal Medicine

B. Imaging

C. Endoscopy - The Gold Standard

• Diagnostic and potentially therapeutic
• Perform within 24 hours of presentation after adequate fluid resuscitation in most patients with overt bleeding
• Earlier (within 12 hours) for high-risk patients (haemodynamic instability, suspected varices)
• Allows Forrest classification of peptic ulcer lesions

• Injection therapy: adrenaline (epinephrine) 1:10,000; tissue adhesives (cyanoacrylate for varices)
• Mechanical: hemoclips, band ligation (varices and Mallory-Weiss)
• Thermal contact: electrocoagulation, thermocoagulation (heater probe)
• Thermal non-contact: argon plasma coagulation (APC), laser photocoagulation

D. Scoring Systems

• Albumin <3.0 g/dL
• INR >1.5
• Mental status altered
• SBP <90 mmHg
• Age >65 years Score ≥2 associated with higher mortality, longer stay, greater cost

---

4. DIFFERENTIAL DIAGNOSIS - COMPLETE ANATOMICAL CLASSIFICATION

ESOPHAGEAL CAUSES

1. Oesophageal varices - portal hypertension; most common variceal source; accounts for ~22% of severe UGIB overall; variceal bleeding is cause of UGIB in cirrhotic patients 59% of the time
2. Gastro-oesophageal varices type 1 (GOV1) - extend along lesser curve
3. Gastro-oesophageal varices type 2 (GOV2) - extend along greater curve toward fundus

---

GASTRIC CAUSES

---

DUODENAL CAUSES

1. Duodenal ulcer (peptic) - most common single cause; posterior DU erodes gastroduodenal artery causing catastrophic haemorrhage
2. Duodenal erosions/duodenitis - NSAIDs, H. pylori, Crohn's disease
3. Duodenal Dieulafoy lesion - less common than gastric; affects 2nd/3rd part of duodenum
4. Duodenal angiodysplasia/Angioectasia
5. Duodenal carcinoma - periampullary carcinoma
6. Ampullary carcinoma
7. Duodenal GIST/stromal tumour

---

SPECIAL/RARE CAUSES

---

5. FREQUENCY SUMMARY (UCLA CURE Database - Sleisenger & Fordtran's)

---

6. KEY CLINICAL PEARLS FOR THE CONSULTANT

1. Melena is an upper GI bleed until proved otherwise - even without hematemesis, melena in a patient <50 is more likely upper GI than lower.
2. 15% of hematochezia has an upper GI source - any patient with hematochezia + haemodynamic instability needs urgent OGD.
3. Always ask about aortic surgery - an aortoenteric fistula classically presents with a "herald bleed" followed by exsanguination. This diagnosis cannot be missed.
4. BUN:Cr ratio ≥30 is a biochemical pointer to UGIB from intraluminal digestion of haemoglobin.
5. Initial haematocrit is unreliable - acute haemorrhage has not had time to equilibrate; haematocrit falls only after haemodilution with IV fluids or extravascular fluid shift.
6. Beta-blockers mask tachycardia; antihypertensives mask hypotension - do not use vital signs alone in these patients to estimate haemorrhage severity.
7. Blatchford score 0 - patient may be safely managed as outpatient with early elective OGD.
8. Forrest Ia/Ib/IIa - require endoscopic hemostasis (combination therapy preferred: injection + mechanical or thermal).
9. PPI infusion - 80 mg bolus then 8 mg/hr continuous infusion for 72 hours post-endoscopic therapy in high-risk ulcers; reduces rebleeding and need for surgery.
10. Octreotide/terlipressin - splanchnic vasoconstriction for suspected variceal bleeding; start empirically while awaiting endoscopy (octreotide: 50 mcg bolus then 25-50 mcg/hr infusion; terlipressin: 2 mg IV 6-hourly).
11. Transfusion threshold - Hb ≤7 g/dL in most patients; ≤9 g/dL in elderly or those with cardiac comorbidities. Restrictive transfusion policy (Hb 7-8 target) is associated with better outcomes than liberal policy in peptic ulcer bleeding.
12. H. pylori testing - test all patients with peptic ulcer bleeding (rapid urease test at endoscopy or stool antigen); eradication reduces 1-year rebleeding rate from ~33% to <5%.
13. Prokinetics before endoscopy - erythromycin 250 mg IV 20-90 minutes before endoscopy improves visualisation significantly; should be used routinely when blood is suspected in the stomach.
14. Rising lactate in hospital = clear predictor of in-hospital mortality; monitor dynamically.

---