In a prospective, controlled, randomized study of the prophylaxis of stress bleeding, 100 ventilated high-risk patients in a surgical intensive care unit received, on a daily basis, 1 g of sucralfate suspension (n = 50) every four hours, or an antacid (n = 50) every two hours. The mean duration of the treatment was about six days in both of the groups. Gastric pH was determined every eight hours. Bleeding was defined as macroscopically visible bleeding. The intragastric pH was less than 4 significantly more often in patients treated with sucralfate. In each group, one case of macroscopically visible bleeding occurred. Both of the patients had a very high risk of bleeding. None of the bleedings influenced the outcome of the patients. When patients with primary thoracic trauma or pneumonia were excluded, nosocomial pneumonia developed in significantly fewer (p less than 0.05) patients in the sucralfate group (three of 29) than in the antacid group (11 of 32). In four of the latter patients, pneumonia influenced the outcome of the patients. Sucralfate provides adequate protection against stress bleeding while also minimizing the danger of pneumonia caused by infection via the gastropulmonary route.