Development of a scoring system to predict mortality from upper gastrointestinal bleeding.

Despite the widespread application of endoscopy in acute upper gastrointestinal bleeding, there is little evidence of improved survival among those who undergo the procedure. To select high-risk patients who might benefit most from diagnostic and therapeutic endoscopy, the authors developed and validated a scoring system based on prognostic indicators of increased mortality. The scoring system was developed from the best clinical predictors of mortality, determined in a prospective study of consecutive bleeding patients. The model was then tested in a prospective validation phase at three hospitals. Three main factors in the model predict mortality: bleeding, including hematochezia, drop in hematocrit of 5%, short duration of bleeding, absence of melena, and hypotension; liver disease, manifested by prolonged prothrombin time and encephalopathy; and renal disease. Patients determined to be at high risk for death using the scoring system might be candidates for aggressive management and for therapeutic endoscopy.