Case Presentation: A 75 year old woman with past medical history of hypertension, diabetes mellitus, Barrett’s esophagus, and chronic constipation presented to the hospital complaining of one week of dyspnea after an international flight. CT pulmonary angiography revealed a non-occlusive pulmonary embolus and an incidental mass in the gastric fundus. Further imaging was suspicious for a gastrointestinal stromal tumor (GIST). She was placed on therapeutic anticoagulation with brief interruption for endoscopic biopsy of the mass on hospital day 4, which was non-diagnostic. On hospital day 5, her hemoglobin dropped from 15 to 12.9 grams/deciliter, and reached a nadir of 6.9 grams/deciliter on hospital day 8. There was no melena, hematochezia, or endoscopic evidence of active bleeding, nor evidence of hemolysis. The patient remained transfusion dependent, and we pursued further imaging of the abdominopelvic cavity, which did not reveal active bleeding. One week after the initial drop in hemoglobin, the patient had several melanotic stools with minimal change in hemoglobin. She tolerated low-molecular weight heparin without bleeding and was scheduled for interval surgical resection after treatment of her thromboembolism.
Discussion: The absence of melena for seven days after a likely upper GI bleed was atypical and prompted an unnecessary diagnostic workup. Despite a high pretest likelihood of bleeding from the biopsy site, we prematurely dismissed this diagnosis and pursued a costly and anxiety-provoking search for alternative diagnoses. Though a late presentation of melena is uncommon, this case serves as a reminder that the absence of melena is not sufficiently sensitive to rule out upper GI bleeding in the appropriate clinical setting.
Also, the finding of a gastric mass on the patient’s imaging had the potential to drastically alter the management of her pulmonary embolism. Had this mass been malignant, the urgency of the resection, duration of anticoagulation, and choice of anticoagulant would differ, as low-molecular weight heparin is a better anticoagulant in the setting of malignancy.1 Though the mass resembled a GIST on imaging, the non-diagnostic biopsy made it difficult to know this with certainty. While it would be appropriate to anticoagulate for a full three months prior to resection of a benign lesion, delaying resection of a malignant mass would be inappropriate.
Conclusions: In the presence of a likely GIST without a tissue diagnosis, deciding the duration of anticoagulation for co-incident venous thromboembolism is difficult, especially when planning surgical resection. Furthermore, the atypical presentation of melena in this patient complicated medical decision making, as the risk of anticoagulation was not clear until the source of bleeding was identified. It is important to remember that while blood typically acts as a gastrointestinal cathartic, the absence of melena does not preclude a diagnosis of recent luminal bleeding.
To cite this abstract:Sehgal P, Goodman E. An Atypical Presentation of Upper Gastrointestinal Bleeding in the Setting of an Acute Pulmonary Embolism: Lessons in Acute Care Gastroenterology, Anticoagulation, and Overuse of Diagnostic Testing. Abstract published at Hospital Medicine 2015, March 29-April 1, National Harbor, Md. Abstract 531. Journal of Hospital Medicine. 2015; 10 (suppl 2). https://www.shmabstracts.com/abstract/an-atypical-presentation-of-upper-gastrointestinal-bleeding-in-the-setting-of-an-acute-pulmonary-embolism-lessons-in-acute-care-gastroenterology-anticoagulation-and-overuse-of-diagnostic-testing/. Accessed January 19, 2020.