Impact of selective and non-selective media on prevalence and genetic makeup of ESBL/pAmpC-producing Escherichia coli in the broiler production pyramid.

Presence of extended-spectrum β-lactamase (ESBL)- and AmpC β-lactamase (pAmpC)-producing Escherichia coli (ESBL/pAmpC-EC) in humans and animals is alarming due to the associated risks of antibiotic therapy failure. ESBL/pAmpC-EC transmission between the human and animal compartments remains controversial. Using cefotaxime-supplemented (selective) media, we recently showed high sample prevalence of ESBL/pAmpC-EC in an integrated broiler chain [i.e. Parent Stock (PS), offspring broilers and their carcasses]. Here, we used a different approach. In parallel with the selective isolation, samples were processed on non-selective media. E. coli isolates were tested for ESBL/pAmpC-production and those found positive were genotyped. For carcasses, total E. coli were enumerated. This approach enabled us to estimate prevalence at the isolate level, which mirrors ESBL/pAmpC-EC colonisation levels. We showed that although present in many animals, ESBL/pAmpC-EC were overall subdominant to intestinal E. coli, indicating that high sample prevalence is not associated with high levels of resistance in individual hosts. This is a relevant aspect for risk assessments, especially regarding the immediate exposure of farm personnel. An exception was a particularly dominant B2/blaCMY-2 lineage in the gut of imported PS chicks. This predominance obscured presence of latent genotypes, however bias towards particular ESBL/pAmpC-EC genotypes from the selective method or underestimation by the non-selective approach did not occur. At the slaughterhouse, we showed a link between total E. coli and ESBL/pAmpC-EC on carcasses. Mitigation strategies for reducing consumers' exposure should aim at suppressing ESBL/pAmpC-EC in the broiler gut as well as controlling critical points in the processing line.