P3449 - Predictors of Short and Long Term Rebleeding in Patients with Overt Gastrointestinal Bleeding: A Prospective Study

Chantal Rizk, MD¹, Anthony Kerbage, MD², Hani Tamim, PhD¹, Walaa G. El Sheikh, BA¹, Don C. Rockey, MD³, Kassem Barada, MD^1
1American University of Beirut, Beirut, Beyrouth, Lebanon; ²Cleveland Clinic, Cleveland, OH; ³Medical University of South Carolina, Charleston, SC

Introduction: Rebleeding is common in patients with gastrointestinal bleeding (GIB) and is associated with increased morbidity and cost. We examined predictors of rebleeding in patients with overt GIB.

Methods: This was a prospective study of patients admitted to a tertiary care center with acute overt GIB between 2013 and 2023. Patients were followed-up in-hospital and after discharge at 1 month, 1 year and yearly thereafter. Rebleeding was defined by the presence of visualized hematemesis, melena, and/or hematochezia ≥ 24 hours after initial endoscopic evaluation and/or hemostasis and initial stabilization, accompanied by change of vital signs or ≥2 g/dL drop in hemoglobin (Hb). Independent predictors of rebleeding were determined after controlling for relevant confounders.

Results: 797 patients with GIB were followed-up for a median of 57 months. Rebleeding rates were 5.3% (in-hospital), 8.9% (1 month), 16.2% (1 year), and 21.8% (end of follow-up). On bivariate analysis, rebleeding was associated with lower Hb level, receiving blood transfusion (BT), stigmata of recent hemorrhage (SRH) on endoscopy, higher Glasgow-Blatchford Score (GBS), and antithrombotics on admission and discharge (Table 1). Independent predictors of 1-month rebleeding were BT of ≥3 units (HR=1.86; P=0.041), the presence of SRH (HR=1.99; P=0.007) and lowest Hb concentration reached during index hospitalization (HR=0.82; P=0.018), with a lower lowest Hb being predictive of higher rebleeding risk. We developed a rebleeding risk score based on SRH, lowest Hb level, and BT of ≥3 units, which had a satisfactory performance (AUC=0.68), and higher scores correlated with increased rebleeding risk. The model had a slightly better predictive accuracy [AUC=0.63, 95%CI 0.54-0.73), p=0.008] than the GBS (AUC=0.58, p=0.087), and the ABC score (AUC=0.62, p=0.012). By the end of follow-up, SRH was the only predictor of rebleeding (HR=1.61; P=0.003), while being on antiplatelet on admission (HR=0.66; P=0.021) or discharge (HR = 0.63; P=0.026) was protective against rebleeding.

Discussion: We identified risk factors for rebleeding after acute overt GIB and developed a simple and novel rebleeding risk score based on SRH, BT, and lowest Hb reached during index hospitalization. The score performed slightly better than risk models. Antiplatelet therapy appeared to have a protective effect against long-term rebleeding. These findings guide patient management and risk stratification.


Disclosures:


Chantal Rizk, MD¹, Anthony Kerbage, MD², Hani Tamim, PhD¹, Walaa G. El Sheikh, BA¹, Don C. Rockey, MD³, Kassem Barada, MD¹. P3449 - Predictors of Short and Long Term Rebleeding in Patients with Overt Gastrointestinal Bleeding: A Prospective Study, ACG 2023 Annual Scientific Meeting Abstracts. Vancouver, BC, Canada: American College of Gastroenterology.