Gastrointestinal hemorrhage in the cancer patient.

The usefulness of fiberoptic endoscopy, performed to identify the cause(s) of gastrointestinal bleeding in cancer patients was assessed by (1) identifying the clinical presentation of bleeding, (2) evaluating the safety and diagnostic yield of fiberoptic endoscopy, (3) determining the frequency distribution of causes of bleeding, and (4) evaluating the clinical course following acute gastrointestinal bleeding in a cancer patient population. Hematemesis, melena, and hematochezia were observed in decreasing order of frequency as manifestations of bleeding. Of 187 endoscopic procedures performed on 133 patients, 75% were bleeding from benign lesions with the majority due to gastric ulceration, gastritis, or duodenal ulceration. One third of patients with tumors involving the gastrointestinal tract were bleeding from another source. Mortality from major hemorrhage was 8%; 55% of patients were alive at the end of 2 years. Endoscopy was performed without complications and contributed to medical management, angiographic therapy, and surgical planning.