IJPPEK GASTROINTESTINAL hemorrhage poses a particularly difficult problein in the patient with leukemia. A general-i7ed bleeding tendency is so common that the special significance of hematemesis may not be recognized. T h e patient often is too sick to permit definitive therapy, should the bleeding site be identified. It is gerierally recognized that the leukemic patient is particularly susceptible to universal erosive gastritis and that specific leukemic infiltrations 01 the stomach, although rare, may on occasion become ulcerated and bleed.1-5, 7-9 It is well to recall that the patient with leukemia also lalls heir to all oE the other bleeding lesions of man. Some of these are curable by surgery or by other manipiilativc means. There appears to be no congncing evidence that the incidence of gastric a n d duodenal ulcer is altered by the presence of leukemia, except as the therapeutic uqe of the steroid hormones may encourage the development of ulcer.
There are reported here, briefly, the cases of five personally observed patients with massive upper gaytrointestinal hemorrhage who were found to have both leukemia and gastric or duodenal ulcer. Patients who had been treated with steroid hormones are not included; two additional leukemic patients with upper gastrointestinal hemorrhage have been observed with bleeding gastric ulcer, but in these the clinical facts suggested that the ul- cers had been induced by ACTH therapy. In all of the cases reported herc thc presence of ulcer was proved, either at operation or at autopsy. T h e patients were studied after adoption at this hospital of the "vigorous diagnostic approach"6 to uuper gastrointestinal hemorrhage, so that the maximum effort was expended in determining the actual sotirce of 1 )Teed ing.