Time-dependent efficacy of seven pharmacological regimens for ulcer healing and delayed bleeding prevention after gastric endoscopic submucosal dissection: A frequentist network meta-analysis

Background Gastric ESD induces large mucosal defects with potential delayed complications, including bleeding and perforation. The comparative efficacy of pharmacological regimens—PPI, P-CAB, H2RA, MP, and their combinations—remains unclear. Methods We conducted a frequentist network meta-analysis of 31 randomized controlled trials evaluating seven pharmacological strategies for post-ESD ulcer healing and delayed bleeding prevention. Outcomes were assessed at 4 and 8 weeks for mucosal healing and ulcer reduction, and across the treatment course for delayed bleeding/perforation. Odds ratios (ORs), mean differences (MDs), and P-scores were calculated. Heterogeneity, inconsistency, sensitivity, and publication bias were assessed. Transitivity was evaluated qualitatively, and certainty of evidence was rated using the GRADE framework. Results At 4 weeks, the point estimate for P-CAB monotherapy favored PPI (OR=1.61, 95%CI 0.96–2.68, P = 0.069), but the confidence interval included the null value and the difference was not statistically significant; P-CAB had the highest numerical P-score ranking; combination regimens (PPI + MP, P-CAB + MP) also yielded numerically better healing outcomes. At 8 weeks, PPI + MP had the highest numerical P-score ranking, though no regimen achieved statistically superior healing compared with PPI. For delayed bleeding and perforation, all regimens demonstrated comparable safety profiles, while H2RA-based treatments showed a non-significant trend toward higher complication risk. Most pairwise comparisons were supported by low-to-moderate certainty evidence. Conclusions While P-CAB-based regimens and MP combinations showed numerically favorable trends for early post-ESD ulcer healing, all comparative findings are limited by mostly low-to-moderate GRADE evidence certainty, and no regimen exhibited statistically significant superiority for ulcer healing. No regimen significantly reduces delayed bleeding or perforation risk.