Late Delayed Bleeding After Gastric ESD with Esomeprazole/Sodium Bicarbonate Versus Conventional Esomeprazole, and Artificial Ulcer Healing with Esomeprazole/Sodium Bicarbonate: A Retrospective Cohort Study

Background/Objectives: Artificial ulcers after gastric endoscopic submucosal dissection (ESD) generally heal within 8 weeks with proton pump inhibitor (PPI) therapy; however, delayed bleeding remains a clinically relevant complication. Esomeprazole/sodium bicarbonate is an immediate-release PPI formulation; its role after ESD is unclear. Methods: This single-center retrospective study included 923 patients who underwent gastric ESD from July 2017 through December 2025. Late delayed bleeding, defined as bleeding after second-look endoscopy, was compared between patients receiving esomeprazole/sodium bicarbonate (n = 532) and conventional esomeprazole (n = 391) using Firth logistic regression adjusted for calendar period and clinical factors. Eight-week artificial ulcer healing was assessed descriptively only in the esomeprazole/sodium bicarbonate group (n = 527). Results: Late delayed bleeding occurred in 26/923 patients (2.8%): 16/532 (3.0%) in the esomeprazole/sodium bicarbonate group and 10/391 (2.6%) in the conventional esomeprazole group. PPI formulation was not associated with late delayed bleeding (adjusted OR 0.872; 95% CI 0.313–2.427). Submucosal fibrosis (adjusted OR 4.502; 95% CI 1.969–10.292), antiplatelet use (adjusted OR 3.620; 95% CI 1.617–8.103), and anticoagulant use (adjusted OR 6.783; 95% CI 1.323–34.770) were independently associated with bleeding. All 26 patients with bleeding were men; sex was therefore excluded from the model. Complete healing was observed in 510/527 patients (96.8%); ulcer diameter ≥ 40 mm was independently associated with incomplete healing (adjusted OR 7.114; 95% CI 2.664–18.998). Conclusions: No statistically significant difference in late delayed bleeding was detected between formulations, although a clinically meaningful difference cannot be excluded. Healing was assessed only in the esomeprazole/sodium bicarbonate group and could not be compared between formulations. Late delayed bleeding was associated with submucosal fibrosis and antithrombotic therapy.