One hundred forty-four critically ill patients admitted to an intensive care setting were randomly assigned to cimetidine or antacid treatament groups. Gastric pH was monitored hourly. One hundred twenty-three (85%) patients demonstrated a fall in pH to less than 4 and were considered to require prophylaxis. Prophylaxis was considered adequate if the measured pH could then be maintained at greater than or equal to 4. Fifty-eight patients received antacids alone, the average requirement being 41 cc/hour. Sixty-five patients received cimetidine. Seventeen (26%) of the cimetidine prophylaxis patients failed to raise their pH and were than placed on hourly administration of antacid with successful elevations of pH to greater than or equal to 4 in all cases on an average supplementary dose of 53 cc/hour. Risk factors, including sepsis, hypotension, head injury, respiratory failure, degree of trauma, and age, were not statistically different in the two treated groups. Using these same criteria, responders to cimetidine could not be differentiated from nonresponders. All patients were protected from significant stress bleeding while on this study. Significant complications of either treatment were minimal. Antacids offered consistent protection against gastric acidity and were 100% effective. A routine schedule of 300 mg every six hours of cimetidine was effective in only 47% of patients, and the maximum dose of cimetidine was effective in only 74% of patients. Hourly measurement of intragastric pH is required for monitoring the response to prophylaxis of stress bleeding in severely ill patients.