Endoscopic evaluation in patients with gastrointestinal bleeding receiving anticoagulation for venous thromboembolism

Gastrointestinal bleeding (GIB) is a common adverse event of anticoagulation in patients undergoing treatment for venous thromboembolism (VTE) and is often assessed by endoscopic evaluation. It remains unclear whether identification of the bleeding source impacts outcomes. Our aim was to describe endoscopic evaluation strategies, management, and outcomes in patients with GIB complicating anticoagulation for VTE.

Patients in the Mayo Clinic VTE registry between March 2013 and October 2021 were prospectively followed. For patients with a first episode of GIB, endoscopic studies and related interventions were reviewed.

Among the 3287 patients, 109 (3%) experienced GIB (median age = 67, 52.3% female). Endoscopic evaluation occurred in 73 patients (67%): upper endoscopy (n = 53, 49%), colonoscopy (n = 41, 38%), and capsule endoscopy (n = 2, 2%). Endoscopy identified a bleeding source in 51 cases (73%), and hemostatic intervention occurred in 19 cases (26%). A bleeding source was not identified in 26 (24%) of cases. Among the 33 (30%) patients who did not undergo endoscopic evaluation, the majority (n = 29, 88%) had an identified source by either prior endoscopy or alternative imaging modality. Anticoagulation was interrupted in 90 (82%) patients for an average of 3.5 days; this was similar regardless of whether a source was identified. During follow-up (median = 103 days), 16 (14.7%) patients experienced a recurrent GIB, and 54 (50%) patients died. Rates of recurrent bleeding, VTE, and death were similar between patients with identified and unidentified bleeding sources.

Although endoscopy has a high diagnostic yield in anticoagulated patients, anticoagulation management and clinical outcomes are similar regardless of GIB source identification. Further studies are needed to determine the net clinical benefit of endoscopy in this high-risk cohort.