# Acute Gastrointestinal Bleeding: An Update and a Practical Diagnostic Approach

## TL;DR

This paper updates the diagnostic approach for acute gastrointestinal bleeding, emphasizing timely evaluation and risk stratification to improve patient outcomes.

## Contribution

The paper provides a practical diagnostic framework and highlights the role of machine learning in identifying low-risk patients.

## Key findings

- Accurate initial evaluation and prompt diagnosis are crucial for managing acute gastrointestinal bleeding.
- Small bowel capsule enteroscopy and device-assisted enteroscopy are key for suspected small bowel bleeding.
- Machine learning models can help identify low-risk outpatients but are not yet widely recommended for general use.

## Abstract

Acute gastrointestinal bleeding (GIB) is one of the most common and dangerous condition in patients admitted in Emergency Departments. The incidence and the mortality of acute GIB remain significant, although some positive trends were observed in recent years. Initial evaluation of GIB needs an accurate assessment of the medical history and the clinical presentation. Physicians should pay attention about the presence of hemorrhagic shock that usually requires urgent diagnosis and treatment. Only a prompt diagnostic approach can identify the source of bleeding and improve the outcomes in acute GIB patients. Risk stratification and time of endoscopy are fundamental issues in the management of upper and lower GIB. Small bowel capsule enteroscopy (SBCE) and device-assisted enteroscopy (DAE) are the basic approaches to suspected small bowel bleeding. Machine Learning Prognostic Models have been proposed, such as alternative prognostic tools in GIB, but they are currently recommended only to identify low-risk outpatients.