Usefulness of the blood urea nitrogen/creatinine ratio in gastrointestinalbleeding

This study was conducted to evaluate the blood urea nitrogen/creatinine (BU N/Cr) ratio for distinguishing an upper versus lower source of gastrointest inal (GI) bleeding. Charts of patients who presented to the emergency depar tment (ED) with the diagnosis of Gt bleeding from August 1995 to August 199 6 were retrospectively reviewed for source of bleeding, initial BUN, Cr, BU N/Cr ratio, hematocrit (Hct), and need for transfusion. A total of 124 pati ents were eligible for inclusion, 71 (57%) of whom were male. A total of 63 (51%) presented with blood in stool and 53 (43%) with bloody emesis; 8 (6% ) had blood in both emesis and stool. A total of 31 (25%) patients had a lo wer GI bleed, 88 (70%) had an upper, and 5 (4%) had both upper and lower bl eeding sources. The mean BUN level was 24 mg/dL, the mean Cr level 1.03 mg/ dL, end the mean BUN/Cr ratio was 24. The mean hemoglobin (Hb) level was 11 .3 g/dL, the mean Hct was 32 g/dL, and 51% required transfusion. Upper GI b leeding was significantly correlated with age younger than 50 (P =.01) and male gender (P=.01; odds ratio, 3.13). Taking into account age and gender, the BUN/Cr ratio correlated significantly with an upper GI source of bleedi ng (P =.03), with a ratio greater than 36 having a sensitivity of 90% and a specificity of 27%. The area under the receiver operating characteristic c urve using age, gender, and BUN/Cr ratio was .73 (95% confidence interval, .62 to .84). Copyright (C) 1999 by W.B. Saunders Company.