To assess the efficacy of cimetidine versus Mylanta II in the prevention of stress erosions, 44 patients at risk were randomized to receive cimetidine 300 mg/6 h intravenously to a maximum of 400 mg/4 h; or Mylanta II 30 ml/h through a nasogastric tube to a maximum of 90 ml/h. The minimum dose of medication was used to maintain hourly gastric pH greater than or equal to 4. All patients were to be endoscoped after 72 h and erosions graded by an endoscopist who had no knowledge of their treatment regime. Grade 3 or 4 erosions occurred in five of 21 cimetidine-treated patients and eight of 16 antacid-treated patients (p greater than 0.05). A gastric pH greater than or equal to 4 was maintained 79.5% of the time by cimetidine and 97.9% of the time by Mylanta II (p less than 0.001). Cimetidine and antacids are equal in the prevention of stress erosions although Mylanta II is superior in hourly pH control. Hourly pH control does not entirely explain the beneficial effect of cimetidine in the prevention of stress ulcers. There was no significant bleeding in this study. Fatalities in patients at risk of developing stress ulcers result from the underlying disease, not from hemorrhage from stress-induced mucosal lesions. In critically ill patients, endoscopic examination should be restricted to the rare case with manifest hemorrhage.