Risk stratification in upper and upper and lower GI bleeding: Which scores should we use?

Area of Science:

• Gastroenterology and Clinical Medicine
• Medical Informatics and Predictive Analytics

Background:

• Risk assessment scores are crucial for managing upper gastrointestinal bleeding (UGIB) but existing tools like the Rockall score, pre-endoscopy Rockall score, Glasgow-Blatchford score, and AIMS-65 have limitations in predicting all critical outcomes.
• There is a notable scarcity of validated risk scores for lower gastrointestinal bleeding (LGIB), necessitating the development of new tools to identify low-risk patients.

Purpose of the Study:

• To review and compare the performance of established and novel risk assessment scores for both UGIB and LGIB.
• To evaluate the quality of external validation for various risk scores used in gastrointestinal bleeding.
• To discuss the comparative efficacy of LGIB scores against UGIB scores (Rockall, Glasgow-Blatchford, AIMS-65) in predicting outcomes in LGIB.

Main Methods:

• Systematic review and comparative analysis of existing literature on risk assessment scores for UGIB and LGIB.
• Evaluation of the performance and external validation quality of scores including Rockall, Glasgow-Blatchford, AIMS-65, Progetto Nazionale Emorragia Digestiva, CANUKA, BLEED, NOBLADS, Strate, Sengupta, and Oakland.
• Discussion of the predictive capabilities of these scores for various clinical outcomes, including severe bleeding, death, and suitability for outpatient management.

Main Results:

• The performance of existing UGIB scores varies, with no single score accurately predicting all important clinical outcomes.
• External validation quality differs significantly across the reviewed UGIB scores.
• Several new scores (BLEED, NOBLADS, Strate, Sengupta, Oakland) have emerged for LGIB, aiming to predict severe outcomes or identify low-risk patients, but their comparative performance requires further investigation.

Conclusions:

• Current risk assessment scores for UGIB have limitations, underscoring the need for more accurate predictive tools.
• The development of validated risk scores for LGIB is an active area, with recent advancements focusing on identifying patients suitable for outpatient care.
• Further comparative studies are essential to establish the optimal risk scores for managing both UGIB and LGIB effectively.