Research trends in post-ERCP pancreatitis: a dual-database bibliometric analysis

Endoscopic retrograde cholangiopancreatography (ERCP) is associated with several adverse events, of which post-ERCP pancreatitis (PEP) is the most frequent despite established preventive measures. Clinical reviews evaluate risk factors and interventions, but the organization and development of the research field have received less attention. This study examined the development of the literature on PEP incidence, risk stratification, prediction, and prevention.

The Web of Science Core Collection (WoSCC) and Scopus were last searched on July 31, 2026. English-language articles and reviews were merged, deduplicated, screened, and manually standardized. The merged dataset was used to analyze publication output, countries, institutions, authors, source journals, and author keywords; citation-dependent analyses were restricted to WoSCC records. VOSviewer and CiteSpace were used for collaboration, co-occurrence, co-citation, clustering, temporal, and citation-burst analyses.

The final dataset comprised 1,686 publications, including 1,414 articles and 272 reviews; 1,449 WoSCC records were available for citation-dependent analyses. Annual publication was sporadic before 1996 and peaked at 133 records in 2025. The United States contributed 485 publications, followed by China (306) and Japan (217), and maintained links with countries across all three collaboration clusters. Gastrointestinal Endoscopy published the most records (n = 170) and was the most frequently co-cited journal (n = 11,065). Temporal analyses showed increasing attention to multimodal prevention and risk assessment. At the retrieval cutoff, adverse events and risk factors had the strongest ongoing keyword bursts, while artificial intelligence and machine learning remained emerging topics.

PEP research has moved toward multimodal prevention and individualized risk estimation. The literature remains concentrated in specialist endoscopy journals and established clinical networks, with substantial contributions from East Asia. Broader multicenter collaboration and independent validation are needed before emerging prediction models can inform routine practice.