Thrombocytopenia and gastrointestinal hemorrhage in the cancer patient: prevalence of unmasked lesions.

Some coagulation deficiencies are known to cause bleeding by unmasking existing gastrointestinal pathology, as opposed to directly causing mucosal blood loss. Characteristics and etiology of gastrointestinal hemorrhage associated with thrombocytopenia have not been analyzed. Our objectives were to correlate the distribution and cause of gastrointestinal bleeding, as diagnosed by fiberoptic endoscopy, with the severity of thrombocytopenia. One hundred thirty-three patients were divided into three groups, determined by platelet count at the time of bleeding (group A: less than 20,000/mm3; group B: 20,000 to 40,000/mm3; group C greater than 40,000/mm3). Results of 187 endoscopies revealed unifocal sources of blood loss in over 50% of each group, and diffuse mucosal oozing independent of gastrointestinal pathology was seen in only 1% of group C. The only significant difference (p = 0.04) comparing unifocal, multifocal, and diffuse sources of bleeding was observed between groups A and C, where in the distribution of multifocal or diffuse sources of bleeding was more common in group A. Esophagitis was more common and gastric ulceration less common in group A. No endoscopic complications occurred. Gastrointestinal bleeding associated with thrombocytopenia is most commonly due to co-existent gastrointestinal pathology as opposed to diffuse mucosal bleeding. Even when an inflammatory process, such as esophagitis or gastritis, affects a particular organ, bleeding is usually unifocal or multifocal as opposed to diffuse even in the presence of moderately severe thrombocytopenia.