Metallo-β-lactamase resistance in Enterobacteriaceae is an artefact of currently utilized antimicrobial susceptibility testing methods

Abstract

Abstract Background MBLs are a major contributor to β-lactam resistance when tested using CAMHB. Despite in vitro resistance, positive outcomes have been reported in MBL-infected patients following carbapenem treatment. The impact of physiological zinc concentrations on this in vitro–in vivo MBL discordance warrants investigation. Objectives To evaluate meropenem in vitro activity against MBL-producing Enterobacteriaceae in zinc-depleted broth (Chelex-CAMHB, EDTA-CAMHB) and assess meropenem efficacy in murine infection models. Methods Neutropenic mice received a meropenem human-simulated regimen of 2 g q8h or levofloxacin 750 mg q24h (for model validation). Zinc concentrations were determined in conventional CAMHB, zinc-depleted CAMHB and epithelial lining fluid (ELF) of lung-infected mice. Results All MBL-producing isolates (NDM, n = 25; VIM, n = 3; IMP, n = 2) examined were meropenem resistant in CAMHB and susceptible in zinc-depleted CAMHB (5- to 11-fold reduction), with zinc depletion having no impact on levofloxacin MICs. Zinc concentrations (mean ± SD) in CAMHB were 0.959 ± 0.038 mg/L and in both zinc-depleted CAMHB and ELF were <0.002 mg/L. In vivo, levofloxacin displayed predictable efficacy consistent with its phenotypic profile, while meropenem produced >1 log unit bacterial killing despite in vitro resistance in conventional CAMHB. Conclusions Results indicate that meropenem in vivo efficacy is best represented by the pharmacodynamic profile generated using MICs determined in zinc-depleted media for MBL-producing Enterobacteriaceae. These translational data suggest that the use of conventional CAMHB for MBL susceptibility testing is inappropriate in distinguishing meaningful in vivo resistance given that zinc concentrations are supraphysiological in conventional CAMHB and negligible at infection sites.