In 62 consecutive patients, 99mTc RBC studies were reviewed and the scintigraphic findings were characterized to determine the ability of 99mTc RBC scintigraphy to estimate bleeding rates in patients with active gastrointestinal hemorrhage. Of the 62 scans, 10 showed early positivity (less than 1 hr) with a strongly intense bleeding focus relative to hepatic activity (group 1), four showed delayed positivity (greater than 1 hr) with a strong focus (group 2), five showed early positivity with a weak focus (group 3), 12 showed delayed positivity with a weak focus (group 4), and 31 studies were negative (group 5). The mean bleeding rates were then calculated by dividing the blood transfusion volume requirements by the duration of active bleeding for each patient, as recorded in the patients' charts. The mean bleeding rates for these groups were 0.4, 0.3, 0.2, 0.15, and 0.1 ml/min, respectively. Group 1 patients had the greatest mean blood-volume loss, longest duration of active bleeding, and highest mean bleeding rate, and they required more aggressive therapy than the other groups. The minimum mean bleeding rate detectable by scintigraphy was 0.1 ml/min. Our results suggest that 99mTc RBC scintigraphy can estimate bleeding rates and identify those patients who are at higher risk for developing massive gastrointestinal hemorrhage requiring more aggressive therapy.