Use of technetium-labeled red blood cell scintigraphy in the detection and management of gastrointestinal hemorrhage.

The effectiveness of technetium 99m-labeled red blood cell scintigraphy in localizing hemorrhage, directing surgical intervention, and screening patients for arteriography was determined in 103 patients. The radionuclide scan result was compared to the bleeding site determined by arteriography, endoscopy, or surgery. Eighty-five patients had a bleeding site identified; 18 patients did not and were excluded. Thirty-one scans were performed in 29 patients for upper gastrointestinal hemorrhage. Five positive scans incorrectly localized an upper gastrointestinal bleeding site, although two scans localized the site, for a scan sensitivity of 8%. Fifty-nine scans were performed in 56 patients with lower gastrointestinal bleeding. Fifteen scans were positive, three incorrectly localizing the hemorrhage. Seventy-four percent of the patients with lower gastrointestinal hemorrhage documented by arteriography, endoscopy, or surgery had negative scans for bleeding. The radionuclide scan sensitivity for lower gastrointestinal bleeding was 23%. Surgery was required in 18 patients for bleeding, 11 of whom had negative scans for bleeding. In seven surgical patients with positive scans, in no instance did the scan direct the surgical intervention. Eighteen patients underwent scintigraphy and arteriography; nearly one half of the patients with negative scans for bleeding had positive localizing arteriograms, although almost one half of the patients with positive scans for bleeding had negative arteriograms. Scintigraphy failed to localize hemorrhage in 85% of the patients. Technetium 99m-labeled red blood cell scintigraphy did not direct surgical intervention, nor did it adequately screen patients for arteriography.