Antibiotics
Network Analysis of Antimicrobial MIC Profiles Reveals Extensive Covariation in Chicken-Associated Escherichia coli from Large-Scale Breeding Flocks in Central Hungary
By Kerek Á, Tornyos GÁ, Szarka E and 4 others.
Abstract
Background
The surveillance of non-clinical bacteria from food-producing animals can reveal antimicrobial selection patterns without conditioning sampling on clinical disease, but categorical resistance estimates are difficult to compare when organism-specific breakpoints are incomplete. We aimed to characterize antimicrobial MIC distributions in chicken-associated Staphylococcus spp., Enterococcus spp., and Escherichia coli from Central Hungary and to determine whether E. coli MIC profiles formed structured covariance networks.
Methods
The dataset comprised 133 isolates collected in 2022-2023 (28 Staphylococcus spp., 59 Enterococcus spp., and 46 E. coli). MIC distributions were summarized using MIC50/MIC90, and E. coli MICs were analyzed by Spearman correlation with Benjamini-Hochberg correction. Historical threshold classification was retained only as a sensitivity analysis, and nine E. coli genomes provided contextual genomic information. Results:Staphylococcus spp. showed lower amoxicillin MIC50/MIC90 values (0.25/1 µg/mL) than Enterococcus spp. (1/256 µg/mL), illustrating marked cross-genus differences. In E. coli, MIC50/MIC90 values included 16/512 µg/mL for ceftriaxone, 128/256 µg/mL for neomycin, 32/128 µg/mL for enrofloxacin, and 32/>512 µg/mL for trimethoprim-sulfamethoxazole. Ten associations met |ρ| ≥ 0.40 and adjusted q < 0.05, forming two connected components that persisted after sensitivity analyses for upper-bound accumulation and the amoxicillin-amoxicillin/clavulanic acid relationship. The nine genomes showed heterogeneous resistance-associated gene calls.