A multi-method framework for antimicrobial resistance hazard ranking: Application to the poultry processing chain

Antimicrobial resistance (AMR) in food-producing animals poses a significant One Health challenge and requires risk-based approaches that extend beyond hazard detection towards actionable prioritization. This study assessed AMR-associated risk along the broiler processing line using three complementary approaches. Neck skin and cecal samples were collected at three successive processing stages: pre-evisceration, post-evisceration and post-chilling. First, microbial loads were quantified by enumerating Campylobacter spp., Escherichia coli and total aerobic bacteria. Second, Salmonella Infantis and extended-spectrum β-lactamase (ESBL) E. coli isolates underwent WGS typing and were risk-ranked based on the number of antimicrobial resistance genes (ARGs), mobile genetic elements (MGEs), virulence genes and/or stress-related genes. Third, all samples underwent shotgun sequencing, and resistome risk ranking of samples was performed based on the prevalence of ARG-carrying contigs, MGEs, presumed host pathogenicity and virulence genes. The proposed AMR-associated hazard ranking scores were able to differentiate isolates and complex samples with contrasting (meta)genomic risk profiles within this dataset. Bacterial counts and the abundance of major ARGs in neck skin samples decreased significantly along the processing line, with the magnitude of this reduction further influenced by flock- and slaughterhouse-specific effects. Overall, this study establishes a semi-quantitative, science-based framework for AMR hazard ranking, which is based on (meta)genomic characterization of both complex samples and their constituent bacterial isolates. This adaptable approach provides a proof-of-concept framework to support AMR hazard prioritization and decision-making with potential for adaptation to other settings.