What is the scoring system for Upper Gastrointestinal (UGI) bleeding?

Scoring Systems for Upper Gastrointestinal Bleeding

Primary Recommendation: Glasgow Blatchford Score

The Glasgow Blatchford Score (GBS) is the preferred scoring system for upper GI bleeding because it has superior sensitivity (99%) for identifying high-risk patients and is specifically designed to determine which patients need clinical intervention or hospitalization. 1, 2

Components of the Glasgow Blatchford Score

The GBS uses only pre-endoscopic clinical and laboratory data, making it immediately available in the emergency department 1:

• Blood urea nitrogen level 1
• Hemoglobin level 1
• Systolic blood pressure 1
• Pulse rate 1
• Presence of melena 1
• Presence of syncope 1
• Evidence of hepatic disease 1
• Evidence of cardiac failure 1

Clinical Application of GBS

• A GBS score of ≤1 identifies patients at very low risk for rebleeding or mortality who may not require hospitalization or inpatient endoscopy, with 98.6% sensitivity for detecting high-risk patients 1, 2
• GBS outperforms other scoring systems in identifying patients who need hospital-based intervention (blood transfusion, endoscopic therapy, or surgery), with an area under the curve of 0.93 3
• GBS accurately identifies low-risk patients suitable for outpatient management, reducing unnecessary hospitalizations without compromising safety 1, 3

Secondary Scoring System: Rockall Score

The Rockall Score exists in two forms and serves different purposes than GBS 1, 2:

Pre-endoscopic Rockall Score

• Uses only clinical variables (age, shock parameters, comorbidities) before endoscopy 1
• Has lower sensitivity (93-96%) compared to GBS, potentially misclassifying 4-7% of high-risk patients as low risk 2
• International consensus groups could not make a firm recommendation for or against using pre-endoscopic Rockall to identify low-risk patients due to insufficient evidence quality 1

Complete Rockall Score

• Requires endoscopic findings in addition to clinical variables, including diagnosis and stigmata of recent hemorrhage 1, 4
• Better discriminative ability for predicting mortality than rebleeding, making it more useful for prognostication after endoscopy 1, 2
• A complete Rockall score <3 indicates excellent prognosis with very low rebleeding and mortality risk 2
• A score ≥6 indicates increased rebleeding risk (18.6% vs 2.9% for score <6) 5

Scoring System NOT Recommended: AIMS65

The AIMS65 score should NOT be used to identify low-risk patients suitable for discharge because it was designed to identify high-risk patients for death rather than low-risk patients for safe discharge 1, 2:

• Lower sensitivity (78-82%) for identifying high-risk patients compared to GBS 2
• Misclassifies approximately 20% of high-risk patients as low risk even at low cutoff values 2
• International consensus recommends against using AIMS65 for identifying patients who may not require hospitalization 1

The Forrest Classification System

The Forrest classification is an endoscopic scoring system that categorizes bleeding peptic ulcers based on their appearance and directly guides endoscopic management decisions 5:

Forrest Classification Categories

• Forrest Ia (spurting arterial bleeding): Highest rebleeding risk (55-59%), requires immediate dual-modality endoscopic hemostasis 5
• Forrest Ib (oozing bleeding): Active bleeding requiring endoscopic therapy 5
• Forrest IIa (visible vessel): High rebleeding risk (43-55%), requires endoscopic hemostasis 5
• Forrest IIb (adherent clot): Requires vigorous irrigation for at least 5 minutes to expose underlying stigmata 5
• Forrest IIc (flat pigmented spot): Low rebleeding risk (5-10%), medical therapy only 5
• Forrest III (clean base): Very low rebleeding risk (5%), medical therapy only 5

Clinical Utility of Forrest Classification

• Forrest classification guides endoscopic intervention decisions, with Forrest Ia, Ib, and IIa requiring mandatory dual-modality hemostasis (mechanical therapy plus epinephrine injection) 5
• Dual-modality therapy reduces rebleeding (OR 0.19,95% CI 0.07-0.52) and need for surgery (OR 0.10,95% CI 0.01-0.50) compared to monotherapy 5
• Forrest classification should be combined with GBS for comprehensive risk stratification, where GBS determines admission need and Forrest classification guides endoscopic management 5

Practical Implementation Algorithm

Step 1: Emergency Department Assessment

• Calculate GBS immediately using available clinical and laboratory data 1, 2
• GBS ≤1: Consider outpatient management with early outpatient endoscopy 1, 2
• GBS 2-6: Admit for early inpatient endoscopy within 24 hours 5
• GBS ≥7: Urgent inpatient endoscopy within 12 hours 5

Step 2: Post-Endoscopy Risk Stratification

• Apply Forrest classification during endoscopy to guide hemostasis decisions 5
• Calculate complete Rockall score for mortality prediction and discharge planning 2, 4
• Complete Rockall <3: Excellent prognosis, consider early discharge 2

Step 3: Contextual Factors for Discharge Decisions

• Consider access to hospital or ambulance services, particularly in rural settings 2
• Assess availability of out-of-hours endoscopy for potential rebleeding 2
• Evaluate patient preferences and social support before discharge 2

Common Pitfalls to Avoid

• Never use clinical judgment alone without a standardized scoring system, as this leads to inconsistent risk assessment and potential missed high-risk patients 2
• Do not rely on AIMS65 for discharge decisions, as it misses too many high-risk patients 1, 2
• Avoid using pre-endoscopic Rockall score as the primary tool for identifying low-risk patients, as it has inferior sensitivity compared to GBS 1, 2
• Never use epinephrine injection alone for endoscopic hemostasis in high-risk Forrest lesions; always combine with mechanical or thermal therapy 5