This systematic review and meta-analysis assessed the effectiveness of fibrin glue, with or without polyglycolic acid (PGA) sheets, in preventing post-ESD bleeding for gastric tumors. Across thirteen studies with 2,728 patients, fibrin-based therapy significantly reduced overall and delayed bleeding compared with standard care. The benefit was primarily driven by the combination of PGA sheets and fibrin glue, which consistently outperformed fibrin alone across bleeding outcomes, including symptomatic events. In contrast, fibrin alone did not demonstrate a consistent protective effect. No significant reduction was observed in acute bleeding, and perforation rates were similar between groups. Although effect estimates favored the intervention at both 4- and 8-week follow-up, no significant interaction by duration was detected, indicating a stable protective effect during the early post-ESD healing phase.

The biological rationale for using PGA sheets and fibrin glue after gastric ESD is grounded in the concept of tissue shielding and enhanced hemostatic stabilization. ESD creates a large artificial ulcer with exposed submucosal vessels that remain vulnerable to acid, bile reflux, and mechanical irritation, predisposing to delayed bleeding. PGA sheets act as a bioabsorbable mechanical barrier that physically covers the ulcer bed, protecting exposed vessels from chemical and mechanical insults while promoting granulation tissue formation and re-epithelialization [27]. Fibrin glue, composed of fibrinogen and thrombin, mimics the final step of the coagulation cascade, forming a stable fibrin clot that adheres to tissue surfaces, seals microvessels, and reinforces hemostasis at the ulcer base [29, 31]. However, fibrin alone may be susceptible to early degradation in the gastric environment. The combined application of PGA sheets with fibrin glue provides synergistic benefit; the sheet offers structural support and sustained coverage, while fibrin acts as both an adhesive and a biologic sealant that secures the patch to the ulcer surface. This dual mechanism: mechanical shielding plus biologically enhanced clot stabilization, likely explains the superior reduction in delayed bleeding observed with combination therapy compared with fibrin alone.

Post-endoscopic submucosal dissection bleeding is commonly classified by timing and clinical presentation into acute, delayed, and symptomatic events, which reflect distinct pathophysiological processes and clinical implications [33, 34]. Acute bleeding typically occurs within the first 24–48 h after ESD and is often related to intra-procedural vessel injury or insufficient immediate hemostasis, whereas delayed bleeding arises days to weeks post-procedure and is influenced by ongoing ulcer exposure to gastric acid, mechanical stress, and the resumption of antithrombotic therapy as coagulum stabilizes and systemic hemostatic balance shifts post-discharge [35]. Delayed bleeding events may also be associated with ulcer sloughing and the breakdown of early clot formation at the resection site [34, 36].

These mechanisms partly explain why Tissue-shielding strategies, particularly PGA sheets combined with fibrin glue, are more effective in preventing delayed rather than acute bleeding after ESD. By providing sustained ulcer coverage during the vulnerable healing phase, they stabilize hemostasis, protect against acid and mechanical stress, and reduce bleeding risk, consistent with our meta-analysis findings.

Compared with the earlier meta-analysis by Li et al. [37], which included four studies (212 PGA-treated vs. 208 controls) and demonstrated a significant reduction in overall post-ESD bleeding (RR 0.33, 95% CI: 0.16–0.69), our study expands and refines the available evidence. While Li et al. focused exclusively on PGA sheets in high-risk early gastric cancer populations, particularly patients receiving antithrombotic therapy, our meta-analysis incorporates a substantially larger and more contemporary dataset, including recent multicenter RCTs and propensity-matched studies. Consistent with their findings, we observed a significant reduction in overall and delayed bleeding, particularly with PGA plus fibrin combinations. However, our analysis further distinguishes between acute, delayed, and symptomatic bleeding and demonstrates that the protective effect is primarily driven by delayed bleeding reduction. Moreover, unlike the earlier study, we show that fibrin alone does not consistently confer benefit, thereby clarifying that the mechanical shielding component appears central to efficacy. These findings strengthen and extend prior conclusions by providing broader generalizability and more granular outcome assessment.

This meta-analysis has several limitations that should be considered when interpreting the findings. First, although randomized controlled trials were included, the majority of studies were retrospective and single-center in design, which increases susceptibility to selection bias, confounding, and variability in clinical practice. Second, heterogeneity was moderate in several bleeding outcomes, likely reflecting differences in patient populations, lesion characteristics, antithrombotic management, operator expertise, and definitions of bleeding events. Third, outcomes were variably reported per patient, lesion, or ulcer, potentially introducing unit-of-analysis inconsistencies also the relatively small number of included studies limited the ability to perform meta-regression analyses to assess the potential impact of baseline covariates on outcomes. Follow-up durations were relatively short (mostly four to eight weeks), limiting assessment of longer-term complications or rebleeding risk. In addition, subgroup analyses particularly for symptomatic bleeding were based on a limited number of studies and were occasionally sensitive to single-study effects. Cost considerations and procedural time were also inconsistently reported, precluding formal evaluation of cost-effectiveness.

Furthermore, substantial procedural heterogeneity existed across studies. Variations in PGA sheet delivery (e.g., forceps vs. clip-and-pull), supplementary clipping, and fibrin glue application (source, volume, and technique) may influence mucosal shield quality, representing a limitation of our pooled analysis.

Future research should prioritize large, well-designed multicenter randomized trials with standardized bleeding definitions and uniform follow-up protocols. Comparative effectiveness research evaluating PGA plus fibrin versus other preventive strategies, alongside cost-effectiveness analyses, will further clarify the optimal approach for routine clinical practice.