Antimicrobial Resistance Gene Profiles in Salmonella spp., Escherichia coli, and Enterococcus spp. from Conventional and Antibiotic-Free Chicken Meat

Antimicrobial resistance (AMR) is a global public health concern, and animal food production contributes to its spread. This study compared antimicrobial resistance gene (ARG) profiles in Salmonella spp., Escherichia coli, and Enterococcus spp. isolated from chicken meat produced under conventional (CONV) and antibiotic-free (ATB-Free) systems using whole-genome sequencing of 207 phenotypically selected isolates. We hypothesized that the ATB-Free production would be associated with fewer resistance genes, but the observed patterns differed among the bacterial groups. In E. coli (41 genes; 61.0% shared between chains), the gene load per isolate was similar between the production systems (median of four in both chains). In Salmonella spp., only five ATB-Free isolates were available, precluding a meaningful comparison between the production systems; a rarefaction analysis showed that the lower gene count observed in this group was consistent with the sampling effort. In Enterococcus spp., the CONV isolates carried a higher gene load (median of two vs. one) and a higher proportion of enzymatic determinants (48.4% vs. 11.8%). This difference was concentrated in isolates carrying the ionophore resistance genes narA and narB (CONV 41.9% vs. ATB-Free 5.9%; p < 0.001), which were not associated with transferable vancomycin or high-level aminoglycoside resistance determinants in this dataset. The resistance determinants associated with the clinically important antimicrobial classes, including blaCTX-M-8, quinolone resistance genes, and fosfomycin resistance genes, were detected in isolates from both production systems. Overall, the observed ARG profiles differed among the bacterial groups, indicating that the relationship between the production system and the resistome was not uniform. Since the sequenced isolates constituted a phenotypically selected subset, the observed ARG frequencies should not be interpreted as population-level prevalence estimates for either production system.