Site of Gastrointestinal Bleeding in Patients on Direct Oral Anticoagulants Versus Aspirin Monotherapy: A Single-Centre Retrospective Cohort Study

Background: Gastrointestinal (GI) bleeding is a recognised complication of long-term antithrombotic therapy. Whereas the overall bleeding risk associated with direct oral anticoagulants (DOACs) and low-dose aspirin has been extensively characterised, the anatomical distribution of bleeding lesions between these two pharmacological classes remains incompletely defined. Methods: We retrospectively analysed all consecutive adult patients admitted to the Emergency Department of San Carlo di Nancy Hospital (Rome, Italy) between January 2022 and March 2025 with a clinical diagnosis of acute GI bleeding. Patients were categorised as DOAC monotherapy or aspirin (ASA) monotherapy at presentation. The primary endpoint was the site of bleeding (upper vs. lower GI tract) by drug group. Secondary endpoints included site of bleeding by specific DOAC molecule and indicators of clinical severity (haemoglobin at admission, transfusion requirement, melena, hematochezia). Logistic and linear regression models were adjusted for age and sex. Because the number of transfused units is count data with a right-skewed distribution, the corresponding linear-regression estimate is reported as an exploratory approximation and interpreted with caution. As adjustment was limited to age and sex, all adjusted estimates are associative rather than causal and do not represent an independent effect of drug class. Results: The merged cohort comprised 316 patients; the head-to-head analysis included 179 patients on monotherapy (DOAC n = 100, ASA n = 79). Upper GI bleeding (UGIB) was equally distributed between groups (55% vs. 54%; OR 1.02, 95% CI 0.57–1.85; p = 1.00), whereas identified lower GI bleeding (LGIB) was numerically less frequent in DOAC monotherapy (34% vs. 49%; OR 0.53, 95% CI 0.29–0.97; p = 0.047), an association that was attenuated and no longer significant after adjustment for age and sex (p = 0.076). Restricting to patients who underwent oesophagogastroduodenoscopy (EGDS), the diagnostic yield was substantially lower in DOAC monotherapy (72% vs. 91%; adjusted OR 0.25, 95% CI 0.08–0.79; p = 0.018). DOAC-treated patients presented with lower haemoglobin (7.9 ± 2.4 vs. 9.3 ± 2.9 g/dL; adjusted β = −1.32 g/dL, p = 0.002) and required transfusion more frequently (68% vs. 49%; adjusted OR 2.50, 95% CI 1.29–4.83; p = 0.007). No statistically significant differences were detected among the four DOAC molecules for site of bleeding or specific lesion, although this analysis was substantially underpowered. Conclusions: In a real-world emergency cohort, DOAC and ASA monotherapy were associated with a comparable proportion of upper GI bleeding, while identified LGIB was numerically less frequent in DOAC users. DOAC users presented with lower haemoglobin and more frequently required transfusion, despite a lower diagnostic yield at EGDS. Among the four DOAC molecules, no significant differences emerged. In sensitivity analyses, the lower identified-LGIB frequency and the higher transfusion burden in DOAC users were attenuated and no longer statistically significant when the analysis was restricted to fully evaluated patients or when count-based and comorbidity-adjusted models were applied. Further studies are warranted to clarify the anatomical distribution of bleeding sources in DOAC users with negative initial endoscopy.