Key points are not available for this paper at this time.
BACKGROUND Accurate classification of adverse events (AEs) in gastrointestinal endoscopy is essential for safety monitoring and quality improvement. The American Society for Gastrointestinal Endoscopy (ASGE) lexicon is widely used, while the classification for AEs in gastrointestinal endoscopy (AGREE) is a recently proposed alternative aiming for broader applicability. AIM To compare the agreement and correlation between the AGREE and ASGE classification systems using real-world data from a Latin American academic endoscopy unit. METHODS A retrospective analysis of a prospective registry was conducted at a tertiary center in Chile, encompassing all endoscopy-related AEs from 2009 to 2022. Each AE was independently graded using both ASGE and AGREE classification systems by two blinded reviewers per system. Interobserver agreement was calculated using Cohen’s Kappa, and inter-scale correlation was assessed using Spearman’s rank test. RESULTS Of 176655 procedures performed, 235 AEs (0.13%) were included. Most events were related to therapeutic procedures, and the most common AEs were cardiorespiratory (42.1%), bleeding (20.9%), and perforation (17.0%). The ASGE system identified 42.1% of cases as incidents and 57.9% as AEs (Kappa = 0.83). AGREE classified 46.0% as non-AEs and 54.0% as AEs (Kappa = 0.74). A strong correlation between both systems was observed (ρ = 0.89; P < 0.001). CONCLUSION The AGREE classification strongly correlates with the ASGE lexicon but excludes more cases as non-AEs and shows slightly lower interobserver agreement. These findings support AGREE as a feasible alternative for AE grading in gastrointestinal endoscopy, particularly in diverse clinical environments.
Corsi et al. (Wed,) studied this question.