Thirty-two long-term ventilated patients were randomly selected for a study of the efficacy of sucralfate (1 g six times per day via gastric tube) versus ranitidine (six 50-mg to six 100-mg doses per day intravenously) for the prevention of upper gastrointestinal bleeding. The patients of the two treatment groups (each 16) were comparable with respect to diseases precipitating acute respiratory failure and risk factors of bleeding, e.g., renal failure, thrombopenia, coagulopathy, and anticoagulant treatment. Mean duration of mechanical ventilation was 7.4 in sucralfate- and 7.7 days in ranitidine-treated patients. During mechanical ventilation, macroscopically visible bleeding developed in three of the sucralfate-treated (18.7 percent) and seven of the ranitidine-treated (43.7 percent) patients. Until the end of the study, only three of the sucralfate-treated but nine of the ranitidine-treated (56.2 percent) patients bled; the difference between the two treatment groups was at all times significant (p less than 0.05). Packed red blood cells had to be administered to the three bleeding patients in the sucralfate group and to seven bleeding in the ranitidine group. Therefore it seems that sucralfate prevented mostly minor bleeding. The high bleeding rate during ranitidine treatment was presumably due to the high number of pH-nonresponders, as almost 30 percent of the gastric aspirates of this group had a pH less than 5. During treatment no difference was found in positive blood culture specimens and bronchial secretions between the two groups. However, nosocomial pneumonia developed in two ranitidine-treated patients, whereas that complication developed in none of the sucralfate-treated patients. In long-term ventilated patients, sucralfate prevented minor upper gastrointestinal bleeding significantly better than ranitidine. However, this does not imply that major upper gastrointestinal bleeding can be prevented by either sucralfate or ranitidine in these patients.