Nonsteroidal anti-inflammatory drug (NSAID)associated injury to the gastrointestinal tract is now a well-established fact. Indeed, studies have shown a prevalence of peptic ulcer in 20 –25% of patients taking these agents. The incidence of gastrointestinal bleeding, perforation, or obstruction in these patients is 1–3%, indicating that approximately one out of every 10 NSAID-associated peptic ulcers becomes complicated (ie, development of hemorrhage, perforation, or obstruction). These complications occur most often in well-defined, high-risk patients, primarily those who are elderly, those who have a history of gastrointestinal disease, and those who are receiving high-dose or multiple-drug regimens. The evaluation of any mucosal-protective agent is extremely difficult. For example, what end-point should be measured to determine efficacy: hemorrhages and erosions in healthy volunteers, peptic ulcers in volunteers or patients, or complication rates in patients taking NSAIDs? It is now generally agreed by most investigators that hemorrhages seen in short-term healthy volunteer studies are of little use. In a study from our laboratory evaluating the ability of the H2-receptor antagonist cimetidine to prevent naproxen-induced gastric injury, it was shown that the antagonist prevented hemorrhages but had no effect on the occurrence of erosions or peptic ulcers. This finding suggests that hemorrhages are acid dependent and are unrelated to more important lesions. Erosions, on the other hand, are probably more significant lesions in short-term healthy volunteer studies, and trials evaluating the prevention of these lesions by various agents are often predictive of findings in arthritic patients during long-term studies. Erosions may therefore be useful as a positive screening mechanism for new prophylactic agents. Healthy volunteer studies have shown that sucralfate is generally ineffective in preventing erosions and ulcers, that H2-receptor antagonists prevent erosions and ulcers in the duodenum but not in the stomach, and that misoprostol prevents both gastric and duodenal erosions and ulcers. –7 Longer-term studies in arthritic patients using ulcer as the end point have demonstrated essentially the same findings. –13 The use of ulcer as an end-point in these longer-term patient studies is justified by the fact that the major complications of NSAID use are those associated with peptic ulcer disease; erosions are found in a high percentage of patients taking NSAIDs chronically, but they are not associated with complications. Studies evaluating sucralfate have shown that it is essentially ineffective in preventing any type of NSAID-associated peptic ulcer disease. Two large, multicenter studies evaluating the H2-receptor antagonist ranitidine, 150 mg twice daily, in the prevention of NSAID-associated peptic ulcer have shown that this agent is significantly better than placebo in the prevention of duodenal ulcer, whereas it has no effect on the prevention of gastric ulcer. However, a recent study has shown that the H2-receptor antagonist famotidine, in addition to preventing duodenal ulcer, significantly reduces the incidence of gastric ulcer when given at high dose in patients receiving chronic NSAID therapy, when compared with placebo. Finally, numerous studies in arthritic patients have shown that misoprostol prevents both gastric and duodenal ulcers in patients taking NSAIDs, –12 although the side-effects caused by misoprostol sometimes prevent its use.