The mortality in patients with upper gastrointestinal bleeding has not changed in the past quarter century in spite of the introduction of new modes of therapy and treatment. In this review we address the possible reasons for a lack of change in mortality and the implications raised for the use of new techniques. We review the factors that affect the mortality of acute upper gastrointestinal hemorrhage and the diagnostic accuracy of upper gastrointestinal endoscopy. Based on this information, we present guidelines for the therapy of the major causes of upper gastrointestinal bleeding. These guidelines should be useful until new therapies have been assessed and become generally available.