• All patients presenting with suspected upper GI bleeding
• Identify patients safe for outpatient management without endoscopy
• Predict need for clinical intervention (transfusion, endoscopy, surgery)
• Triage urgency of inpatient endoscopy

• BUN: Elevated BUN reflects absorbed blood — the highest-impact variable
• Haemoglobin: Differentiates by sex (male threshold higher)
• Systolic BP: Defines haemodynamic compromise
• Heart rate, melaena, syncope, liver disease, cardiac failure: Binary clinical factors

• Score 0: Very low risk. Outpatient management safe. Sensitivity 98.6% for identifying low-risk.
• Score 1–3: Low risk. Consider outpatient endoscopy.
• Score 4–7: Moderate risk. Admit. Endoscopy within 24h.
• Score ≥ 8: High risk. Urgent endoscopy. ICU-level monitoring.

• GBS: Pre-endoscopy. Predicts NEED for intervention. Best for triage.
• Rockall Score: Post-endoscopy. Predicts REBLEEDING and mortality. Complements GBS.
• AIM65: Also pre-endoscopy. Uses albumin, INR, mental status, SBP, age > 65. Comparable performance to GBS.

• GBS 0: Discharge. Oral PPI. Outpatient GI follow-up within 72h.
• GBS 1–3: Consider admission for observation vs outpatient endoscopy based on clinical context.
• GBS 4–7: Admit. Cross-match 2–4 units PRBCs. IV PPI infusion. Endoscopy within 24h.
• GBS ≥ 8: Resuscitate. Target Hb 70–80 g/L. Urgent endoscopy (< 12h if haemodynamically unstable). GI/surgery consult.

Developed by Blatchford, Murray, and Blatchford at Glasgow Royal Infirmary in 2000 as a pre-endoscopy triage tool to safely identify patients who did not require urgent hospitalization. Its inclusion of purely clinical and biochemical variables (no endoscopic findings required) allows application at the point of first assessment in any setting.