D. A. Hill1, L. Khoury1, M. Kopp1, M. Panzo1, T. Bajaj1, C. Schell1, A. Corrigan1, R. Rodriguez1, S. Cohn1 1Northwell Health At Staten Island University Hospital,Department Of Surgery,Staten Island, NEW YORK, USA
Objective: With the advent of proton pump inhibitors and H. Pylori treatment, the old dogma “the most common cause of lower gastrointestinal (GI) bleeding is upper GI bleeding” may no longer be accurate. Similarly the 1994 publication by McGuire describing the correlation between a transfusion requirement of > 4 units of packed red blood cells (PRBCs) requiring operative intervention 60% of the time may also no longer be valid. We sought to determine the most common causes of GI bleeding in patients without an obvious source and the likelihood of transfusion, and endoscopic or surgical intervention.
Method: We queried our hospital database for GI hemorrhage during 2015, excluding patients with obvious upper GI source (hematemesis), or anal pathology. We collected data from patients with GI bleeding defined as bright red blood per rectum, melena or a positive fecal occult blood test. The primary endpoints were etiology of GI bleed, amount of transfusions required and operative, endoscopic or angiographic interventions performed.
Results: 93 patients were admitted with GI bleeding: mean age was 74 years old (range 19 to 95), 52% were male, and mean hemoglobin was 8.2 (range 3.5 to 14.2). 74% received blood transfusions with an average of 2 units transfused per patient (range 0 to 9); 22% received ≥ 3 units of packed red blood cells. The etiology of bleeding was: 17% upper GI source; 15% lower GI source; and in 68% the source remained unknown. Bleeding stopped spontaneously in 86% of patients and 9% died (deaths were related to refusal of transfusions, requests for comfort care measures or withdrawal of care). 71% underwent inpatient endoscopy with only 6% undergoing a therapeutic endoscopic intervention (none of those undergoing intervention were transfused ≥ 3 units). No patient had surgical or interventional radiologic procedures related to their GI bleed.
Conclusion: Gastrointestinal bleeding, without an obvious source on presentation, rarely requires operative or interventional radiologic intervention. Blood transfusions were not predictive of the need for therapeutic endoscopic intervention which was required in only 6% of patients.