Validated Scoring Tools for UGIB Risk Stratification
Use the Glasgow-Blatchford Score (GBS) as your primary tool to identify low-risk patients who can be safely discharged without hospitalization or urgent endoscopy, and use the complete Rockall score after endoscopy to predict rebleeding and mortality risk. 1
Primary Recommendation: Glasgow-Blatchford Score
The GBS is the preferred pre-endoscopy tool for initial risk stratification in acute UGIB 1. This score incorporates:
- Hemoglobin level
- Blood urea nitrogen
- Pulse rate
- Systolic blood pressure
- Presence of syncope or melena
- Evidence of hepatic disease or cardiac failure
A GBS ≤1 identifies very low-risk patients who can be safely managed as outpatients without early endoscopy 2. The 2019 international consensus guidelines specifically recommend this threshold for discharge decisions, as these patients have minimal risk for clinical intervention or mortality 2.
The GBS demonstrates superior performance compared to pre-endoscopy Rockall scores for identifying patients requiring hospital-based intervention 1. Recent validation studies confirm GBS outperforms other pre-endoscopy scores in predicting need for blood transfusion (AUC 0.82-0.83) 3, 4.
Secondary Tool: Complete Rockall Score
After endoscopy, use the complete Rockall score to predict rebleeding and mortality risk 1. This score adds endoscopic variables to clinical parameters:
- Age
- Shock status
- Comorbidities
- Endoscopic diagnosis
- Stigmata of recent hemorrhage (active bleeding, visible vessel, adherent clot)
The complete Rockall score has been validated across multiple countries and demonstrates better discriminative ability for mortality than for rebleeding (though it performs adequately for both) 1.
Alternative Consideration: AIMS65 for Mortality Prediction
If your primary concern is in-hospital mortality prediction, AIMS65 may be considered 2. This simpler score includes:
- Albumin <3.0 g/dL
- INR >1.5
- Altered mental status
- Systolic BP <90 mmHg
- Age >65 years
AIMS65 demonstrates superior performance for mortality prediction (AUC 0.71-0.80) compared to GBS 5, 4, and outperforms GBS for predicting ICU admission 4. However, the 2019 international guidelines specifically recommend against using AIMS65 to identify low-risk patients for discharge (100% consensus vote) 2, as it performs poorly for this purpose.
Practical Algorithm
Step 1 - Emergency Department Assessment:
- Calculate GBS immediately using readily available clinical and laboratory data
- GBS ≤1: Consider outpatient management with urgent follow-up endoscopy
- GBS >1: Admit for inpatient management and endoscopy within 24 hours
Step 2 - Post-Endoscopy Risk Stratification:
- Calculate complete Rockall score incorporating endoscopic findings
- Use this to guide intensity of monitoring, timing of discharge, and rebleeding risk counseling
Step 3 - Mortality Risk Assessment (if needed):
- If specific mortality risk stratification is required for ICU triage or family discussions, calculate AIMS65
- AIMS65 ≥2 indicates significantly elevated mortality risk
Important Caveats
Do not delay endoscopy to calculate scores. These tools inform triage decisions but should not postpone therapeutic intervention in unstable patients 1.
Endoscopic stigmata trump clinical scores for high-risk lesions. Active arterial bleeding, visible vessels, or large ulcers (>2 cm) on posterior duodenal or lesser gastric curvature walls require endoscopic therapy regardless of clinical score 1.
Score performance varies by outcome. No single score optimally predicts all outcomes 5, 3. GBS excels at identifying need for intervention and transfusion; Rockall predicts mortality better than rebleeding; AIMS65 is best for mortality and ICU needs. Match the score to your clinical question.
Newer scores (ABC, HOLD-B) show promise in recent studies 6, 7 but lack the extensive validation and guideline endorsement of GBS and Rockall. Stick with established tools until prospective multicenter validation confirms superiority.