The combined data from 16 prospective trials (2133 patients) appear to suggest that antacids prevent stress ulcer bleeding more effectively than does cimetidine. However, the use of occult blood detection methods to diagnose stress ulcer bleeding may have led to the recognition of clinically insignificant bleeding. When the data from these trials are categorized according to the criteria used for the diagnoses of bleeding (either occult blood detection or clinically overt bleeding), there was no significant difference between antacids and cimetidine in the prevention of overt bleeding (3.3% of 458 compared with 2.7% of 402 patients who bled, respectively; p = 0.69). In addition, both agents were more effective (p less than 0.001) than placebo (15% of 720 patients who bled) in the prevention of overt bleeding. Cimetidine and antacids are equal in preventing significant stress ulcer bleeding.