Diagnostic efficacy of computed tomography angiography for acute gastrointestinal bleeding

Computed tomography angiography (CTA) is the radiologic test for acute gastrointestinal bleeding (GIB). We aimed to (1) identify the clinical factors associated with (a) positive CTA findings, (b) positive esophagogastroduodenoscopy (EGD) findings after a negative CTA, and (2) compare bleeding severity on EGD by CTA findings. Using data from a retrospective cohort of patients, we examined the factors associated with positive CTA findings and those associated with bleeding on EGD after a negative CTA. We examined the severity of upper GIB on EGD by findings on CTA and compared bleeding severity with the Glasgow Blatchford Score (GBS). We also evaluated the factors associated with high-grade bleeding lesions on EGD. A total of 1677 patients were included in this study, of which 229 (13.7 ​%) had positive CTA findings. A history of GIB, hematochezia on presentation and any length of stay in the ICU were associated with positive CTA findings. Among patients with negative CTA results, hematemesis on presentation [OR 2.94; 95 ​% CI (1.53, 5.64)] and high-risk GBS [OR 5.19; 95 ​% CI (2.02, 13.35)] were associated with finding upper GIB on EGD. ICU admission for 4+ days was associated with higher-grade bleeding lesions. Those with positive CTA findings had a higher proportion of Forrest 1 lesions (10.8 ​% vs 3.7 ​%). Our findings support a low threshold for EGD among patients with extended ICU stays (4+ days) even after a negative CTA. The GBS, validated for risk stratification in upper GIB, remains a useful predictive tool even in the context of a negative CTA.