Efficacy of Fibrin Sealants and Polyglycolic Acid Sheets in Reducing Postoperative Hemorrhage Following Gastric Endoscopic Submucosal Dissection: A Meta-analysis of Randomized and Observational Studies.

BACKGROUND: Endoscopic submucosal dissection (ESD) is the standard treatment for early gastric neoplasms with low lymph node metastasis risk but is associated with notable bleeding complications, especially in high-risk patients. This study aims to evaluate the effectiveness of fibrin glue (FG) based strategies, with or without polyglycolic acid (PGA) sheets, in reducing postoperative hemorrhage. METHODS: We performed a systematic search of PubMed, Scopus, and the Web of Science from inception to December 2025. We included randomized controlled trials and observational studies comparing FG or FG plus PGA sheets versus standard care in adult gastric ESD patients. The primary outcome was delayed bleeding. Risk ratios (RRs) with 95% confidence intervals (CIs) were calculated using a random-effects model. RESULTS: Thirteen studies involving an analytical cohort of 2,728 patients were included. The meta-analysis demonstrated that the intervention significantly reduced the risk of delayed bleeding (RR: 0.49; 95% CI: 0.28, 0.84; P = 0.01) and overall bleeding (RR: 0.52; 95% CI: 0.29, 0.91; P = 0.02) compared to standard care. Subgroup analysis revealed that the benefit was primarily driven by the combination of PGA sheets and FG (RR: 0.33; 95% CI: 0.17, 0.63; P < 0.001), while FG alone did not show a statistically significant reduction (RR: 0.62; P = 0.26). Similarly, symptomatic bleeding was significantly reduced only in the PGA + FG subgroup (RR: 0.25; P = 0.009). No significant differences were observed regarding acute bleeding (RR: 0.90; P = 0.81) or procedural perforation (RR: 1.09; P = 0.87). CONCLUSION: The combined application of PGA sheets and fibrin glue significantly reduces delayed hemorrhage after gastric ESD, likely due to a dual mechanism of mechanical shielding and biological stabilization. Fibrin glue alone does not consistently confer a protective benefit.