Pitfalls in genotypic antimicrobial susceptibility testing caused by low expression of bla(KPC) in Escherichia coli

Background: There is a growing interest in the rapid genotypic identification of antimicrobial resistance (AMR). In routine diagnostics, we detected multiple KPC-positive Escherichia coli (KPC-Ec) with discordant phenotypic meropenem susceptibility from a single patient's blood cultures, which prompted a more thorough investigation. Objectives: We investigated the potential clinical relevance of, and the mechanism behind, discordant phenotypic and genotypic meropenem susceptibility in KPC-Ec. Methods: WGS was used to perform a comparative analysis of the isolates' genetic characteristics and their bla(KPC-2) locus. Expression of bla(KPC-2) was determined by quantitative PCR and the potency of meropenem hydrolysis was determined using a semi-quantitative carbapenem inactivation method. An in vivo infection assay using Galleria mellonella was performed to assess the potential clinical relevance of KPC expression in E. coli. Results: Despite the presence of bla(KPC-2), three of five isolates were susceptible to meropenem (MICVITEK2 = 16mg/L). The isolates with high MICs had significantly higher bla(KPC-2) expression, which corresponds to phenotypic meropenem inactivation. The genetic environment of bla(KPC-2), which may impact KPC production, was identical in all isolates. In vivo infection assay with G. mellonella suggested that meropenem was effective in reducing mortality following infection with low-expressing KPC-Ec. Conclusions: Our findings clearly highlight a limitation of genotypic AMR prediction for bla(KPC). For the time being, genotypic AMR prediction requires additional analysis for accurate antibiotic therapy decision-making.