Escherichia coli Antibiotic Susceptibility Patterns for Infants Admitted to NICUs Across the United States from 2009 to 2017

Abstract

Background:Escherichia coli (E. coli) is a leading cause of infections among term and preterm newborn infants. Continued surveillance of neonatal E. coli antibiotic susceptibility patterns is important to optimize empiric antibiotic prescription for infants at risk for infection, in light of evolving reports of multidrug-resistant gram-negative bacteria in all settings. Our objective was to determine E. coli epidemiology and antibiotic susceptibility patterns among a large sample of infants admitted to neonatal intensive care units (NICUs) across the United States from 2009 to 2017. Methods: Retrospective observational study using the Premier Database, including infants born from 2009 to 2017 and admitted to academic or community NICUs contributing microbiology data during the study period. We analyzed antibiotic susceptibilities for E. coli isolated from blood, cerebrospinal fluid, and urine. We focused on clinically relevant and priority susceptibility categories: (1) ampicillin nonsusceptible; (2) aminoglycoside nonsusceptible; (3) carbapenem nonsusceptible; and (4) extended-spectrum β-lactamase (ESBL; phenotypic definition). We determined the proportion of infants with nonsusceptible organisms in each category by year and tested for changes over time. Lastly, we assessed susceptibility patterns by specimen source, birthweight, and timing of infection. Results: Of the 117,484 included infants, 733 (0.6%) had at least 1 E. coli episode, of which 721 (98.4%) had available susceptibility results, from 69 centers. Patient and center characteristics of infants with E. coli are shown in Table 1. Most organisms were tested against ampicillin (99.9%), gentamicin (99.6%), and ceftriaxone (91.5%). Figure 1 shows nonsusceptibility rates for the categories of interest. Overall, ampicillin nonsusceptibility ranged from 63.3% to 68.6% per year (mean, 66.8%±1.5%); aminoglycoside nonsusceptibility ranged from 10.7% to 23.2% (mean, 16.8%±4.5%); carbapenem nonsusceptibility was 0% for all years; and ESBL ranged from 1.2% to 11.3% (mean, 5.1%±3.4%). We detected no statistically significant trends for any of the categories of interest over time (all P > .05), and susceptibility trends were consistent when repeated by specimen source, birthweight, and timing of infection. Conclusions: We found stable, yet concerning, patterns of E. coli antibiotic nonsusceptibility among infants admitted to NICUs across the United States from 2009 to 2017. Rates of ampicillin nonsusceptibility and aminoglycoside nonsusceptibility were higher than previous reports. ESBL E. coli rates were low but present among neonatal patients. No carbapenem nonsusceptible E. coli was identified. These findings can inform empiric antibiotic prescription for infants admitted to NICUs across the United States.

Funding: None

Disclosures: None