Goals: To assess epidemiologic features and predictive factors of mortality of acute upper gastrointestinal bleeding (UGIB). Study: During a 6-month period, a prospective population based study including all the UGB occurring in a geographic area of 3 million people was conducted. Data from cirrhotic patients were compared with those of noncirrhotic patients. Results: A total of 2,133 UGIB were recorded, 21.9%in cirrhotic patients (n = 468). Endoscopic hemostasis was performed in 46.5%and 8.3%in cirrhotic and noncirrhotic patients, respectively (P < 0.001). Mortality during hospitalization was 23.5%in cirrhotic patients and 11.2%in noncirrhotic patients (P < 0.001). Six independent predictive factors of mortality were observed in both patient groups: a prothrombin level less than 40%, an UGIB occurring in inpatients, a concomitant digestive carcinoma, a hematemesis revealing the UGIB, a recent use of steroid drugs, and age over 60 years. Four other predictive factors of mortality were also identified in noncirrhotic patients. Conclusions: Although epidemiologic features, clinical course, management, and prognosis of UGIB were quite different in cirrhotic and noncirrhotic patients, the majority of predictive factors of mortality were the same in both patient groups. These data underline the major role of debilitated status and hepatic failure in the prognosis of UGIB in cirrhotic patients. Goals: To assess epidemiologic features and predictive factors of mortality of acute upper gastrointestinal bleeding (UGIB). Study: During a 6-month period, a prospective population based study including all the UGB occurring in a geographic area of ​​3 million people was conducted. Results: A total of 2,133 UGIB were recorded, 21.9% in cirrhotic patients (n = 468). Endoscopic hemostasis was performed in 46.5% and 8.3% in cirrhotic and noncirrhotic patients, respectively (P <0.001). Mortality during hospitalization was 23.5% in cirrhotic patients and 11.2% in noncirrhotic patients (P <0.001). Six independent predictive factors of mortality were observed in both patient groups: a prothrombin level less than 40%, an UGIB occurring in patients, a concomitant digestive carcinoma, a hematemesis revealing the UGIB, a recent use of steroid drugs, and age over 60 years. Four other predictive factors of mortality were also identified in noncirrhotic patients. Conclusions: Although epidemiologic features, clinical course, management, and prognosis of UGIB were quite different in cirrhotic and noncirrhotic patients, mortality in the same in both patient groups. These data underline the major role of debilitated status and hepatic failure in the prognosis of UGIB in cirrhotic patients.