OBJECTIVE. There is disagreement over the reliability of technetium Tc 99m (99mTc)-labeled erythrocyte scintigraphy in the localization of active lower gastrointestinal hemorrhage. A previous study at The New York Hospital-Cornell Medical Center that showed a superior sensitivity for localization of scintigraphy versus angiography in surgical patients led the authors to emphasize scintigraphy as the diagnostic test of first choice in the clinical diagnostic algorithm. The authors hypothesized that tagged erythrocyte scintigraphy can be used accurately as the primary diagnostic modality in localizing acute bleeding and guiding surgical intervention. METHODS. The authors conducted a 5-year, retrospective analysis of 224 inpatients who underwent scintigraphic imaging for diagnosis and localization of active lower gastrointestinal bleeding. Using scintigraphy as the primary diagnostic test, with colonoscopy, upper endoscopy, and angiography as adjunctive studies, 99mTc-labeled erythrocyte scans were performed at the clinician's discretion and were reviewed again for study purposes by two nuclear radiologists who were blinded to clinical outcome. Adjunctive diagnostic tests also were ordered for clinical indications. RESULTS. Using delayed periodic scintigraphic imaging, results of 115 scans (51.3%) demonstrated bleeding, with 96 scans (42.9%) localizing to a specific anatomic site. Patients with positive scans were five times more likely to require surgery (p < 0.005) than patients with negative scans, and surgical patients were twice as likely to localize by scintigraphy (p < 0.0001). Fifty patients (22.3%) required surgical intervention to control hemorrhage and had a bleeding site confirmed by both clinical and pathologic examinations. Forty-eight of those patients (96%) had a bleeding site determined preoperatively. For 37 patients with bleeding sites localized preoperatively by scintigraphy, 36 (97.3%) had correct localization based on surgical pathology. Only one patient required a subtotal colectomy solely because of nonlocalized bleeding. No patient bled postoperatively, and there was no mortality in either operated or nonoperated patients. The mean volume of transfused erythrocytes was similar in both scan-localized and nonlocalized surgical patients. CONCLUSION. When performed correctly and interpreted conservatively, scintigraphy is a useful and safe means of guiding segmental resection, and should be the primary tool used in the diagnosis of patients with active lower gastrointestinal bleeding.