Clinical features, management and outcomes of gastrointestinal bleeding in patients treated with oral anticoagulants

Abstract

Background: Gastrointestinal (GI) bleeding with the different antithrombotics may present peculiarities in terms of location, precipitating factors, clinical management and prognosis. Purpose(s): Our objective was to compare the profile and clinical course of GI bleeding with direct oral anticoagulants (DOAC) versus vitamin K antagonists (VKA). Method(s): We carried out a retrospective study of all consecutive patients treated in a tertiary hospital during 2018 and 2019, and who met the following selection criteria: 1) diagnosis of confirmed or probable GI bleeding; 2) red blood cell transfusion; 3) treatment with an oral anticoagulant; 4) absence of concomitant antiplatelet therapy. We collected information on comorbidities, bleeding risk scores, baseline treatments, and clinical course of bleeding. We compared adjusted all-cause mortality at 12-months between DOACs and VKAs groups. Result(s): We identified 115 patients with GI bleeding, mean age 83+/-9 years, 63% women, 50.4% on DOACs and 49.6% on VKAs. NOACs group showed more recent anticoagulation history, and more complex clinical profile, with older age (85 vs. 82 years, p=0.026), number of comorbidities (2.7 vs. 2.1, p=0.049), CHA2DS2VASc score (5.2 vs. 4.2, p=0.001) and ORBIT score (3.9 vs. 3.3, p=0.047). There were no differences in the location of bleeding (60.5% lower GI tract), number of units transfused (mean 2.6), or hemoglobin nadir (mean 7.6 g/dL). Notably, 42% of the patients on DOACs were receiving the high dose at the time of bleeding, 63% of them had some risk criterion for overdose (age>80 years, weight <60 kg, moderate P-glycoprotein inhibitors), and 37% had >1 of these criteria. 12-month cumulative mortality was high, but significantly lower in the DOACs group (15.5% vs. 35.7%, adjusted HR 0.31, p=0.002). Conclusion(s): Patients who experience GI bleeding with anticoagulants represent a therapeutic challenge, with both high age and prevalence of comorbidities. One-year mortality is remarkably high, particularly in the VKA group. Our findings emphasize the need for close monitoring and optimization of preventive strategies in this complex clinical scenario.