Efficacy of humanized carbapenem and ceftazidime regimens against Enterobacteriaceae producing OXA-48 carbapenemase in a murine infection model.

Enterobacteriaceae producing the OXA-48 carbapenemase are emerging worldwide, leaving few treatment options. Efficacy has been demonstrated in vivo with ceftazidime against a ceftazidime-susceptible OXA-48 isolate but not with imipenem despite maintaining susceptibility. The relationship between phenotype and in vivo efficacy was assessed for OXA-48 producers using humanized regimens of 2 g doripenem every 8 h (q8h; 4 h infusion), 1 g ertapenem q24h, 2 g ceftazidime q8h (2 h inf), and 500 mg levofloxacin q24h. Each regimen was evaluated over 24 h against an isogenic pair (wild-type and OXA-48 Klebsiella pneumoniae strains) and six clinical OXA-48 isolates with and without other extended-spectrum β-lactamases in immunocompetent and neutropenic murine thigh infection models. Efficacy was determined using the change in bacterial density versus 24-h growth controls in immunocompetent studies and 0-h controls in neutropenic studies. Bacterial reductions of ≥1 log CFU were observed with all agents for the wild-type strain. Consistent with low MICs, ceftazidime and levofloxacin exhibited efficacy against the isogenic OXA-48 strain, whereas doripenem did not, despite having a susceptible MIC; no activity was observed with ertapenem, consistent with a resistant MIC. Similar trends were observed for the clinical isolates evaluated. Ceftazidime, levofloxacin, and ertapenem efficacy against isogenic and clinical OXA-48-producing strains correlated well with phenotypic profiles and pharmacodynamic targets, whereas efficacy with doripenem was variable over the MIC range studied. These data suggest that carbapenems may not be a reliable treatment for treating OXA-48 producers and add to previous observations with KPC and NDM-1 suggesting that genotype may better predict activity of the carbapenems than the phenotypic profile.