Sixty-five surgical ICU patients at high risk of developing acute erosive gastritis and bleeding received prophylactic antacid treatment to maintain a gastric pH of at least 5.0. A similar control group of 61 patients received no specific prophylaxis. All patients in both groups developed microscopic bleeding; however, microscopic bleeding did not influence outcome. In the control group, eight (13.1%) patients developed moderate visible bleeding, as compared to seven (10.8%) patients in the antacid group, an insignificant difference. A single patient in the control group developed severe GI bleeding due to acute erosive gastritis. Antacid prophylaxis did not prevent macroscopic bleeding and there was no correlation between the number of risk factors in individual patients and the rate of upper GI bleeding. We conclude that antacid is not required to prevent upper GI bleeding in high-risk critically ill patients.