(447) Tranexamic Acid in Upper Gastrointestinal Bleeding Is Associated With Increased CVA, MI, & VTE Risk

Introduction: Upper gastrointestinal bleeding (UGIB) is common and life-threatening. Tranexamic acid (TXA) is an antifibrinolytic agent that has shown benefit in traumatic and postpartum hemorrhage. However, evidence for its use in UGIB is conflicting, particularly with regards to vascular adverse effects. A recent trial (HALT-IT) showed no mortality benefit, but increased rates of venous thromboembolism (VTE), without a difference in arterial thrombotic events (myocardial infarction [MI] and cerebrovascular accident [CVA]). We sought to characterize these vascular adverse events utilizing a very large national cohort.

Methods: We queried TriNetX, a database of deidentified, mixed health record and claims-derived data of over 98 million patients from across the U.S. We used ICD-10-CM codes for hematemesis and melena to define patients with a first instance of UGIB between 08/2003 and 07/2023, then defined a cohort that received TXA within a week preceding or following UGIB diagnosis. We then used 1:1 propensity matching across 15 different demographic and comorbidity categories to compare this cohort against UGIB patients who did not receive TXA for rates of MI, CVA, pulmonary embolism (PE) and deep vein thrombosis (DVT) within 1 week of meeting inclusion criteria.

Results: In the timeframe assessed, 1,864,022 patients were diagnosed with UGIB, 5,771 of whom received TXA. After propensity matching, we found the TXA group was significantly associated with all four vascular adverse outcomes, with the most prominent being DVT (OR 2.5, CI 2.1 – 3.1), followed by MI (OR 2.2, CI 1.7 – 2.8), PE (OR 2.1, CI 1.5 – 2.8), and CVA (OR 1.7, CI 1.4 – 2.1).

Conclusions: TXA use was significantly associated with both venous and arterial thrombotic events in the week following UGIB diagnosis. Our findings are limited by the use of deidentified patient data precluding our ability to determine specific timing and dosage of TXA, and lack of acute indicators of illness severity. Nevertheless, the association of increased rates of MI and stroke with TXA that we report is novel and should further caution clinicians considering the use of TXA in this population.