The syndrome of a large drop in hematocrit in hospitalized patients: clinical features and gastrointestinal bleeding outcomes.

Background Hemoglobin/hematocrit decreases in hospitalized patients without evidence of obvious gastrointestinal (GI) bleeding are common and often attributed to GI lesions. We hypothesized that inpatients with hematocrit declines without clinically overt GI bleeding have distinct clinical features and hence aimed to evaluate this clinical syndrome. Methods We evaluated consecutive GI consults for patients with a decrease (drop) in hematocrit. Patients with documented prior GI bleeding, melena, hematochezia, hematemesis, and melenemesis were excluded. Hematocrit drops were calculated from preconsult zenith and nadir values and identified as acute (within 7 days), intermediate (7–30 days), or remote (beyond 30 days) on the basis of the shortest time to a hematocrit drop of at least 6%. Results We identified 101 patients having hematocrit drops without clinical evidence of GI bleeding. Seventy-six patients (75%) had an acute drop in hematocrit with a mean decline of 9.5%, whereas the 25 patients with an intermediate or remote drop in hematocrit had a mean decline of 12.5%. Endoscopic evaluation was performed in 55 patients. Five patients (5%) had a lesion consistent with GI bleeding; identified lesions included ischemic colitis, duodenal ulcer, esophageal varices, and postsphincterotomy ulcer. In the remaining 96 patients, a clinical explanation for the hematocrit drop was identified in 56 patients (58%), most commonly attributed to sepsis/shock, volume shifts, and postprocedure bleeding. Conclusions Hematocrit drops without visible GI bleeding were uncommonly caused by bleeding intestinal lesions, contrary to current dogma. Non-GI explanations for hematocrit drops were often identified, although the cause of hematocrit drop remained unexplained in many patients.