The prevalence and significance of leukocytosis in upper gastrointestinal bleeding.

Although leukocytosis has long been recognized to occur in patients with hemorrhage, there are no data regarding leukocytosis in patients with upper gastrointestinal bleeding. We evaluated the prevalence and significance of the admission white blood cell count in consecutive patients admitted to Grady Memorial Hospital with upper gastrointestinal bleeding seen prospectively over a 50-month period. Any white count greater than 8.5 x 10(3)/mm3 was considered abnormal. Of the 731 patients eligible for the study, leukocytosis was seen in 463 (63%). When compared to patients with a normal white count, patients with leukocytosis on admission were more likely to be tachycardic (31.4% versus 24.3%, P = 0.04) and hypotensive (10.9% versus 5.7%, P = 0.018), required more units of blood (4.6+/-5.9 versus 3.5+/-6.0, P = 0.01), had a longer hospital stay (7.3+/-9.7 versus 5.9+/-6.2 days, P = 0.01), and required more frequent surgical intervention for bleeding (8.0% versus 4.2%, P = 0.04). No significant difference in mortality was seen between patients with leukocytosis and those with a normal white count (8.7% versus 6.4%, P = 0.27). Leukocytosis is common in patients with upper gastrointestinal bleeding, appears to reflect the severity of the bleeding episode, and is associated with a more complicated course.