Phenotypic Definitions for Identifying Carbapenemase-Producing Carbapenem-resistant Enterobacteriaceae

Abstract

Background. Evidence suggests that much of the increase in carbapenem-resistant Enterobacteriaceae (CRE) in the U.S. is due to the spread of carbapenemase-producing (CP) strains. However, resistance mechanism testing is not widely used. A phenotypic definition that reliably identifies CP-CRE could help target prevention.

Methods. Escherichia coli, Enterobacter spp., and Klebsiella spp. isolates that were nonsusceptible to any carbapenem based on local laboratory results were collected from six Emerging Infections Program sites. Isolates underwent susceptibility testing and PCR for the most common U.S. carbapenemases (KPC, NDM, IMP, VIM, OXA-48). The proportion of false positives (FP) (i.e., met phenotypic definition but not CP-CRE), and false negatives (FN) (i.e., did not meet phenotypic definition but was CP-CRE) were calculated for ten phenotypic CRE definitions that included a variety of carbapenem and 3rd and 4thgeneration cephalosporin susceptibility patterns. Phenotypic definitions with FP ≤ 35% and FN ≤ 6% were considered acceptable and further stratified by organism.

Results. Overall, 212 isolates were included, of which 50 (24%), 84 (40%), and 78 (36%) were E. coli, Klebsiella spp. and Enterobacter spp., respectively. Seventy-four (35%) were KPC-CRE (5 E. coli, 57 Klebsiella spp., 12 Enterobacter) and five (2%) were NDM-CRE (5 Klebsiella spp.). The proportion of FP and FN for the ten phenotypic definitions ranged from 17% to 48% and 1% to 11% respectively. Two phenotypic definitions met criteria for further evaluation including: 1). Current CDC phenotypic definition – nonsusceptible to any carbapenem (excluding ertapenem) and resistant to all 3rd generation cephalosporins tested (FN = 4%, FP = 31%) and 2). Resistant to any carbapenem (excluding ertapenem) (FN = 6%, FP =18%). Klebsiella had a higher proportion of FN and lower proportion of FP than E. coli or Enterobacter

Conclusion. No phenotypic definition perfectly identifies CP-CRE; the proportion of FP and FN also might vary by organism. Two phenotypic definitions appeared to have a potentially acceptable proportion of FP and FN and might be useful to target CRE surveillance and prevention efforts; however, testing across a broader group of sites and carbapenemases is needed.

Disclosures. All authors: No reported disclosures.