Primary CT Angiography Vs Colonoscopy in Acute Lower Gastrointestinal Hemorrhage

Background and Aims: Acute lower gastrointestinal bleeding (LGIB) is a common cause of hospitalization and results in significant morbidity and cost. Advancements in computed tomography with angiography (CTA) imaging capability brings into question whether its use as a primary intervention for undifferentiated LGIB in place of colonoscopy could result in improved outcomes. Methods: Retrospective cohort analysis of admitted patients >18 years old who presented between January 2010 and August 2018 with an initial episode of LGIB who underwent colonoscopy or CTA as primary diagnostic intervention. Chi-square or Fisher's exact were used to compare primary outcomes of bleeding detection and intervention rates, t-test for length of stay (LOS) between CTA and colonoscopy. Logistic regression models were built to compare source identification and hemostatic intervention rate. Results: Of 258 total patients, 162 underwent initial colonoscopy vs 96 CTAs. When controlling for hypotension, anticoagulation, transfusions, and time to intervention, colonoscopy was associated with decreased LOS (5.00 vs 6.9, P = 0.001), a higher probability of source identification (OR 3.64, P < 0.001, CL 1.92-6.90) and hemostatic intervention (OR 8.62, P < 0.001, CI 3.54-21.0) compared to CTA. In a subgroup analysis of diverticular bleeds, CTA had higher rates of therapeutic intervention compared to colonoscopy (18% vs 3.8%, P < 0.001, OR 0.09, CI 0.02-0.46) without mortality benefit or shorter LOS. Conclusion: In patients with undifferentiated LGIB, colonoscopy as the primary modality for evaluation results in higher source identification, hemostatic intervention, and shorter LOS. In diverticular bleeds, early CTA was associated with increased hemostatic intervention rate.