Carbapenem and Cephalosporin Resistance among Enterobacteriaceae in Healthcare-Associated Infections, California, USA

We analyzed antimicrobial susceptibility test results reported in healthcare-associated infections by California hospitals during 2014–2017. Approximately 3.2% of Enterobacteriaceae reported in healthcare-associated infections were resistant to carbapenems and 26.9% were resistant to cephalosporins. The proportion of cephalosporin-resistant Escherichia coli increased 7% (risk ratio 1.07, 95% CI 1.04–1.11) per year during 2014–2017.


T he Centers for Disease Control and Prevention (CDC)
identified carbapenem-resistant Enterobacteriaceae (CRE) as an urgent public health threat and extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae as a serious public health threat (1). Antimicrobial-resistant pathogens, such as CRE, can spread across regions when infected or colonized patients transfer between healthcare facilities without infection control measures in place to prevent transmission (2). Therefore, tracking regional changes in antimicrobial resistance (AMR) is essential to inform public health prevention and containment strategies.

The Study
Healthcare-associated infection (HAI) pathogen data reported to the National Healthcare Safety Network (NHSN) can be used to estimate the prevalence of AMR among hospitals within a region (3)(4)(5). Hospitals provide pathogen and antimicrobial susceptibility test results for <3 microorganisms when reporting central line-associated bloodstream infections (CLABSI), surgical site infections (SSI), and catheter-associated urinary tract infections (CAUTI) to NHSN (6). Data on molecular mechanisms of resistance are not collected for CLABSI, SSI, or CAUTI.
We applied CDC definitions to identify antimicrobialresistant phenotypes among Enterobacteriaceae, including Escherichia coli, Klebsiella species, and Enterobacter species, reported in CLABSI, SSI, and CAUTI by general acute-care hospitals in California (3). We included multiple pathogens per HAI if reported. California hospitals report HAI data for <28 surgical procedures; we included pathogen data from any SSI reported. We excluded HAI data reported by other hospital types, such as critical access and long-term acute-care hospitals, due to limited HAI data reported by these hospitals.
According to CDC definitions, CRE were resistant to imipenem, meropenem, doripenem, or ertapenem. Extended-spectrum cephalosporin-resistant (ESCR) Enterobacteriaceae were resistant to ceftriaxone, ceftazidime, cefepime, or cefotaxime. We applied modified phenotype definitions from Magiorakos et al. to identify multidrugresistant (MDR), extensively drug-resistant (XDR), and pandrug-resistant (PDR) Enterobacteriaceae (7). Susceptibility data for 2 antimicrobial drugs (ceftaroline and fosfomycin) included in these definitions were not available in our NHSN data. Resistance was defined by an isolate's nonsusceptibility to >1 agent (e.g., imipenem) within a category of antimicrobial drugs (e.g., carbapenems) and the total number of antimicrobial categories (<15) for which the isolate was nonsusceptible. MDR Enterobacteriaceae were nonsusceptible to >3 antimicrobial categories; XDR Enterobacteriaceae were nonsusceptible to all but 1 or 2 antimicrobial categories, and PDR Enterobacteriaceae were nonsusceptible to all antimicrobial categories. We also assessed the phenotype difficult-to-treat (DTR) proposed by Kadri et al. (8). DTR included an intermediate or resistant result to all reported agents within carbapenem, cephalosporin, and fluoroquinolone categories, as well as piperacillin-tazobactam and aztreonam when results were available.
We used log binomial regression models to estimate statewide, year-to-year change in the proportion of antimicrobial-resistant Enterobacteriaceae during 2014-2017. To understand regional differences in CRE and ESCR Enterobacteriaceae, we performed a subgroup analysis in which we aggregated HAI data in 2-year increments and measured percentage resistance by county when susceptibility test results for >30 Enterobacteriaceae were available. CDC has explored risk adjustment for regional-level comparisons using NHSN data and determined unadjusted measures are satisfactory until additional covariates are adopted in NHSN (9).
We completed data analyses in SAS version 9.4 (SAS, http://www.sas.com) and spatial analyses in ArcMap version 10  We observed decreasing trends in carbapenem resistance (RR 0.90; 95% CI 0.80-1.01) and in the DTR phenotype (RR 0.88; 95% CI 0.77-1.00) among Klebsiella species reported in HAI. Among Enterobacteriaceae assessed for the DTR phenotype, Klebsiella species accounted for 86% (n = 193) of DTR isolates and comprised 23% of the overall total of Enterobacteriaceae analyzed among HAI. In addition, 1 XDR Klebsiella pneumoniae was reported in HAI during 2014-2017 and no PDR Enterobacteriaceae were reported.
Percentages of CRE and ESCR phenotypes varied by county and reporting years ( Table 2; Figures 1, 2). Carbapenem and cephalosporin resistance was higher in California regions more densely populated with hospitals and residents, such as the greater Los Angeles region and San Francisco Bay area. Counties with hospitals reporting <30 Enterobacteriaceae may still have antimicrobial-resistant HAI or receive patients from healthcare facilities where antimicrobial resistance is endemic.
Several factors limit the interpretation of our results. Only 4 years of data were available for measuring AMR trends. Selective reporting of susceptibility test results may have restricted sample sizes and increased the potential for sampling bias to affect our results. Furthermore, there may be differences in how California hospitals and laboratories interpret MIC breakpoints or changes in how breakpoints are applied over time. Data on molecular mechanisms of resistance are not collected in CLABSI, SSI, or CAUTI, which limits our understanding of how transmissible elements, including ESBL and carbapenemases, may contribute to the trends we observed.

Conclusions
Increases in carbapenem, cephalosporin, and MDR E. coli reported in HAI by California hospitals are concerning, given that E. coli are common causes of both hospital and community-associated infections. ESBL-producing E. coli have been reported in community-associated urinary tract infections among patients in California, with estimates of resistance among E. coli from 5% up to 17% in complicated pyelonephritis (10,11). MDR and DTR Enterobacteriaceae further limit treatment options and present  AMR prevention and containment strategies may depend on the local prevalence. For example, prompt detection and rapid, aggressive containment responses to individual AMR cases can be effective in low-prevalence regions. Admission screening and empiric use of transmission-based precautions for patients at high risk for AMR might be more feasible in higher-prevalence regions.
Healthcare facilities can prevent HAI and the spread of AMR by implementing best practices in infection control and antimicrobial stewardship. State and local health departments can coordinate prevention efforts across the healthcare continuum, investigate and control outbreaks in healthcare facilities, and set expectations for healthcare facilities to communicate patients' AMR infection and colonization status during all patient transfers. Decreasing trends in carbapenem resistance and in the DTR phenotype among Klebsiella species, often the focus of AMR containment efforts, indicate the potential effectiveness of such prevention strategies (5). Nonetheless, increases and regional variation in carbapenem-resistant and ESCR E. coli highlight the urgent need for ongoing, local infection prevention and antimicrobial stewardship efforts.
Aspects of this work were supported by Centers for Disease Control and Prevention Epidemiology and Laboratory Capacity grant funds.

About the Author
Mr. Rizzo is an epidemiologist with the Healthcare-Associated Infections Program at the California Department of Public Health. His work focuses on surveillance of antimicrobialresistant healthcare-associated infections and evaluation of prevention programs. ally similar to penicillins, with the substitution of a carbon atom (carba-) for a sulfur atom. This substitution creates a double bond on the pentane ring, which becomes a pentene ring (-penem).
The first carbapenem, thienamycin (theion ["sulfur"] + enamine [an unsaturated compound that forms the backbone of the molecule] + -mycin [suffix for drugs produced by Streptomyces spp.]), was discovered in 1976 in culture broths of the newly recognized species Streptomyces cattleya. Thienamycin rapidly decomposes in the presence of water, which limits its clinical utility.
The first carbapenem approved for use in the United States was imipenem, the stable N-formimidoyl derivative of thienamycin, in 1985. Resistance to imipenem, encoded on a mobile genetic element, was first identified in Pseudomonas aeruginosa in Japan in 1991, and carbapenemase-producing organisms have since spread globally.