The study by Ben-Menachem and colleagues [1] showed no differences among sucralfate, continuous-infusion cimetidine, or placebo in preventing stress-related mucosal damage for patients in a medical intensive care unit. However, several factors may have influenced the cimetidine results in this study, including the use of an inappropriately low initial infusion rate for cimetidine, the significantly lower use of enteral feeding in the cimetidine group, and a significantly higher rate of sepsis on admission to the intensive care unit for recipients of cimetidine. Another potential confounder is that although patients who received omeprazole within 24 hours of admission to the intensive care unit were excluded, the antisecretory effects of omeprazole may last for several days after discontinuation [2].

In addition to these issues, the authors' conclusion that neither cimetidine nor sucralfate should be routinely used for patients in a medical intensive care unit is too broad because they excluded many patients who may have been at an increased risk for gastrointestinal bleeding. For example, of the 877 patients admitted to the medical intensive care unit during the study period, only 301 (34%) were eligible for randomization [1]. The excluded patients met one of many extensive exclusion criteria, such as gastrointestinal hemorrhage within the past year, use of antiulcer drugs within 24 hours of admission to the intensive care unit, use of nonsteroidal anti-inflammatory agents, or closed head injury. Clinical trials of cimetidine in critically ill patients included persons with some of these conditions and showed a clinically relevant reduction in upper gastrointestinal tract bleeding [3, 4]. Even though many of the patients in the study by Ben-Menachem and colleagues had other risk factors for gastrointestinal hemorrhage, the exclusion of essentially two thirds of their patients in the medical intensive care unit precludes such a broad conclusion.