Table 1_Tranexamic acid use and progression to clinically important bleeding in critically ill patients with non-clinically important non-variceal upper gastrointestinal bleeding: a retrospective cohort study.docx

Progression to clinically important GIB did not differ significantly between the TXA and non- TXA groups before matching (5.8% vs. 5.7%; P=0.951, 95% CI: 0.450–1.707) or after matching (5.8% vs. 7.8%; P=0.664, 95% CI: 0.562–3.164); Exploratory univariate conditional logistic regression analysis showed that TXA use was not associated with this outcome (OR 1.333, 95% CI 0.562–3.164; P=0.514). In the matched cohort, TXA was not associated with significant differences in endoscopy requirement, transfusion requirement, hospital length of stay, in-hospital mortality, or thrombotic complications.

Inverse probability of treatment weighting analysis yielded consistent results for the primary outcome (5.1% vs. 5.8%; P=0.418, 95% CI: 0.840–1.522). In critically ill patients with non-clinically important non-variceal UGIB, TXA use was not associated with a statistically detectable reduction in progression to clinically important GIB or improvements in major clinical outcomes. Routine empirical use of TXA in this population should be approached with caution.</p>