Aims
Post-endoscopic submucosal dissection (ESD) bleeding (PEB) is classified as early-onset or late-onset, and controversies regarding risk factors and preventive measures continue. Dialysis patients are widely recognized as a high-risk group for PEB; however, no established criteria exist for managing anticoagulants during dialysis or timing the resumption of antithrombotic therapy after gastric ESD. This study aimed to clarify the frequency and timing of PEB in dialysis patients at a single institution and compare these findings with non-dialysis patients.
Methods
We retrospectively analyzed 47 cases undergoing gastric ESD at our institution between April 11, 2023, and August 19, 2025. Second-look endoscopy (SLE) was performed the following day in all cases. PEB was defined as either (1) requiring emergency endoscopy due to hematemesis or melena, or (2) demonstrating bleeding on endoscopic re-examination with a hemoglobin decrease of ≥2 g/dL. The onset timing was classified as early (Day 0–1) or late (Day ≥2). Cases undergoing prophylactic cauterization or clipping during SLE without evident bleeding, followed by no subsequent bleeding, were classified as the prevent group. PEB was tallied on a per-patient basis.
Results
Of the 47 subjects, 3 were dialysis patients and 44 were non-dialysis patients. Overall PEB occurred in 4/47 cases (8.5%). PEB was observed in 2/3 dialysis patients (66.7%), both of which were late-onset (Day 3, Day 6). Among non-dialysis patients, PEB occurred in 2/44 cases (4.5%) (one case had two events on Day 1 and Day 6, the other on Day 6). A significant difference in PEB occurrence was observed between dialysis and non-dialysis patients (Fisher's exact test, two-tailed p=0.016). The prevent group comprised 9 cases. The dialysis patients who developed PEB were: one case where the anticoagulant during dialysis was changed from nafamostat to dalteparin on Day 3 and hematemesis occurred on the same day; and another case where aspirin was restarted on Day 3, changed from aspirin to prasugrel on Day 5, and hematemesis occurred on Day 6.
| Dialysis Patients(n=3) | Non-Dialysis Patients(n=44) | |
| Sex(Male, %) |
(2, 66.7) |
(31, 70.5) |
| Age(Median, Range) | (69, 66-76) | (76, 41-94) |
| PEB | 2 | 2 |
| PEB Late-Onset | 2(100%) | 1(50%) |
| Prevent Group | 0 | 9 |
|
UML classification (U, M, L) |
(1, 1, 1) | (7, 21, 16) |
|
Tumor Size (>30mm, <30mm) |
(0, 3) | (16, 28) |
| Multiple Tumors | 0 | 5 |
|
Aspirin (continue, suspend) |
(1, 1) | (4, 2) |
|
P2Y12 Inhibitors (continue, suspend) |
(0, 0) | (0, 3) |
|
Direct Oral Anticoagulant (continue, suspend) |
(0, 0) | (0, 7) |
| Cilostazol continue | 0 | 1 |
|
Best-J score (mean) |
4.3 | 1.6 |
Conclusions
In this institution's study, dialysis patients were at high risk for PEB, with delayed onset being particularly prominent. These findings are consistent with the PEB risk in dialysis patients demonstrated in previous studies using the BEST-J score. The anticoagulant used during dialysis and the antiplatelet agent restarted afterward may be implicated in delayed PEB. For dialysis patients after gastric ESD, it is considered appropriate to use nafamostat as the anticoagulant during dialysis and manage antiplatelet therapy cautiously with aspirin monotherapy for approximately one week postoperatively.