Clinical variables associated with positive angiographic localization of lower gastrointestinal bleeding

Background: Localizing the source of severe lower gastrointestinal (GI) bleeding is often difficult but is important to plan the extent of colonic resection. The purpose of the present paper was to audit the Auckland Hospital experience of selective angiography, in localizing lower GI bleeding. Methods: Patients admitted to Auckland Hospital with rectal bleeding and who subsequently had angiography were evaluated by reviewing their clinical notes and radiological results during a 7-year period (1997-2003). Data collected included demographic details, haemodynamic parameters, change in haemoglobin level, requirement of blood transfusion within 24 h before the procedure, site of the bleeding and pathology. Results: The notes of 88 patients (male, n = 51; median age 69 years, range 8-99 years) were available for review and analysis. The site of bleeding was localized in 38 (51%); 30 of them had bleeding in the right colon or small bowel and eight in the left colon. Positive localization correlated with: haemodynamic instability P < 0.0001; drop in haemoglobin level of >= 50 from previous admission (P = 0.02); transfusion requirement of >= 5 units of blood within 24 h (P < 0.0001). Logistic regression analysis showed transfusion requirement of >= 5 units to achieve haemodynamic stability to be the most powerful predictor of accurate localization (odds ratio, 40). Conclusion: Catheter angiography for acute lower GI bleeding will successfully localize a point of bleeding in approximately 50% of patients. The most useful clinical indicator for positive angiography was haemodynamic instability particularly in those who require transfusion of >= 5 units of blood to achieve haemodynamic stability.