A Retrospective Review of Upper Gastrointestinal Bleed Outcomes During Hospital Admission While on Oral Anticoagulation


South Atlantic


Grand Strand Medical Center

Document Type


Publication Date



upper gastrointestinal bleed, ugib, direct-acting oral anticoagulants, apixaban, rivaroxaban, dabigatran, warfarin, ppi, proton-pump inhibitor


Digestive System Diseases | Gastroenterology | Internal Medicine



Direct-acting oral anticoagulants (DOACs) are approved for stroke prevention in non-valvular atrial fibrillation and treatment of venous thromboembolism. Most recent guidelines recommend DOACs over warfarin for most diagnoses given their predictable pharmacodynamics, lack of required monitoring, and safety profile. Specific outcomes such as shock, acute renal failure, and blood transfusion requirement while on oral anticoagulation compared to no anticoagulation remain unknown in patients with upper gastrointestinal (GI) bleeds.


This retrospective study used the HCA Healthcare Enterprise Data Warehouse (EDW) to analyze 13,440 patients aged >18 years that were admitted with an upper GI bleed from January 2017 to December 2019. The patients were categorized based on oral anticoagulant (i.e. rivaroxaban, apixaban, dabigatran and warfarin). The control group was patients admitted with an upper GI bleed not on oral anticoagulation. We evaluated the severity of upper GI bleeds while on oral anticoagulation based on the outcomes: mortality rate, length of stay, acute renal failure, shock, and need for packed red blood cell transfusions (pRBC). Comorbid conditions assessed were coronary artery disease (CAD), chronic obstructive pulmonary disease (COPD), heart failure (HF), atrial fibrillation (AF), venous thromboembolism (VTE), peripheral vascular disease (PVD), tobacco abuse, alcohol abuse, and chronic kidney disease (CKD). Home use of proton pump inhibitors (PPI), aspirin, and P2Y12 inhibitors were also evaluated.


Patients on a DOAC without home PPI have a mortality odds ratio of 3.066 with a confidence interval (CI) greater than 95% (1.48-6.26, p<0.05) compared to patients on a DOAC and home PPI. Patients on warfarin and no home PPI have a mortality odds ratio of 5.55 (95% CI (1.02-30.35), p<0.05) compared to those on warfarin with home PPI use. In the no anticoagulation group, those not on PPI have an odds ratio of 3.28 (95% CI (2.54-4.24), p<0.05) of death compared to home PPI use. There was no statistical difference in mortality between each DOAC and warfarin.

There was no difference in the presence of acute renal failure or shock when comparing each DOAC, warfarin, and no medication. For patients presenting with GI bleed, 0.8414 units of pRBC were transfused. Patients not on oral anticoagulation were found to have statistically significant decrease in pRBC transfusion if they did not report alcohol use, CKD, HF, AF, VTE, PVD. Patients on DOACs and alcohol use have an average pRBC transfusion count that is 0.922 units more than those without reported alcohol use (p=0.006). In the warfarin group, there was no statistical significance noted when comparing pRBC transfusions and also when comparing to baseline comorbidities.


The retrospective study leads us to conclude that overall, patients taking the DOACs or warfarin had no statistically significant increase in RBC transfusions, length of stay, shock, acute renal failure, or mortality rate compared to patients who were not on oral anticoagulation. Home PPI use was shown to lower odds of mortality in patients on anticoagulation who presented with upper GI bleeding. PPI use had no effect on the need for transfusion or length of stay in patients on anticoagulation. These results can help predict which patients are likely to have higher mortality based on the use of home PPIs.

Publisher or Conference