Predicting the outcome of nonvariceal upper GI bleeding: can we guess right?

Abstract

Upper gastrointestinal bleeding (UGIB) represents a common emergency in clinical practice, with an estimated incidence of about 100/100,000 people per year [1]. Traditionally, patients with UGIB are routinely hospitalised, and this results in substantial resource consumption. While there is no doubt that hospitalisation is mandatory for variceal haemorrhage in cirrhotic patients [2], it has become increasingly clear that peptic ulcer bleeding (by far the most common cause of nonvariceal UGIB) is highly inconstant in severity and outcome [3]. In fact, if patients are stratified according to the most commonly used classification of bleeding ulcers [4], the risk of rebleeding varies between 5% in patients with a clean ulcer base and 55% in those with active bleeding at endoscopy, the need for surgical intervention ranges between 0.5 and 35%, and mortality varies between 2 and 13% [3]. Even more importantly, patients in the lowest risk class account for about 40% of the entire population of ulcer bleeders [3], and could be managed as outpatients, leading to substantial cost savings.