Rebleeding from peptic ulcers is a major unsolved problem in the management of acute upper gastrointestinal bleeding. Our goal was to review what is known and what remains to be learned about the effectiveness of antisecretory therapy for acute upper gastrointestinal bleeding. We reviewed the data regarding the effectiveness of endoscopic therapy, the prediction of those at increased risk for rebleeding, and the effectiveness of antisecretory drug therapy in preventing rebleeding with or without endoscopic hemostasis. Proton pump inhibitor therapy without endoscopic hemostasis is ineffective clinically for stopping bleeding or preventing rebleeding. Endoscopic hemostasis remains the cornerstone of therapy. The data are consistent with the notion that reliable maintenance of the intragastric pH at > or = 6 after endoscopic hemostasis is associated with the lowest rebleeding rates. H2-receptor antagonists are ineffective for achieving this goal. Intermittent bolus and oral administration of proton pump inhibitors are equivalent and fail to achieve this goal, which can only be accomplished by bolus administration of a proton pump inhibitor (e.g., 80 mg) followed by a constant infusion (e.g., 8 mg/hr). Whether the combination of endoscopic hemostasis and pH control is equal or superior to selected second-look endoscopy is unknown. A treatment algorithm is suggested.