As of 2016, county-level CRE surveillance was conducted in selected US metropolitan counties at 8 EIP sites (Colorado, Georgia, Maryland, Minnesota, New Mexico, New York, Oregon, Tennessee); surveillance subsequently expanded to 2 additional sites (California in 2017 and Connecticut statewide in 2018) (20). The total population of the 10 participating areas under surveillance in 2020 was an estimated 23.2 million, of which an estimated 4.9 million were children (21).

County-level ESBL-E surveillance started in July 2019 in selected counties at 6 EIP sites (Colorado, Georgia, Maryland, New Mexico, New York, Tennessee). The total population of the 6 participating areas under surveillance in 2020 was an estimated 3.0 million, of which an estimated 626,000 were children (21) (Appendix).

Beginning in 2016, we defined an incident pediatric CRE case as the first isolation during a 30-day period of Klebsiella pneumoniae, K. oxytoca, K. aerogenes, Enterobacter cloacae complex, or Escherichia coli resistant to >1 carbapenem (imipenem, meropenem, doripenem, ertapenem) from a normally sterile body site (Appendix) or urine specimen from a surveillance area resident <18 years of age. We defined an incident pediatric ESBL-E case as the first isolation during a 30-day period of E. coli, K. pneumoniae, K. variicola, or K. oxytoca resistant to >1 extended-spectrum cephalosporin (ceftazidime, cefotaxime, or ceftriaxone) and nonresistant (i.e., susceptible or intermediate) to all tested carbapenems from a normally sterile body site or urine specimen from a surveillance area resident <18 years of age. To prevent duplication with CRE surveillance, we excluded ESBL-E isolates that were carbapenem resistant. For both ESBL-E and CRE, if a new specimen meeting the case definition was collected >30 days after the patient’s last incident case with the same organism, it was also reported as an incident case. CRE and ESBL-E cases were identified through a query of automated testing instruments based on laboratory protocols (Appendix).

All incident CRE cases, as well as all incident ESBL-E cases from a sterile source, underwent medical record review by using a standardized case report form to collect patient demographics, underlying conditions, healthcare exposures and outcomes, location of specimen collection, associated infection types, and antimicrobial susceptibility testing results (20) (Appendix). For ESBL-E cases identified from urine sources, for each year, first incident cases per species in a patient were reviewed.

Cases were considered hospital onset if the incident culture was collected >3 days after hospital admission. All other cases were considered community onset and further classified as either 1) healthcare-associated if the person had hospitalization, surgery, residence in a long-term care facility or long-term care acute care hospital, or chronic dialysis in the year before culture or had an indwelling device in the 2 days before culture; or 2) community-associated if none of those risk factors were identified.

We submitted a convenience sample of isolates from all EIP sites to CDC for confirmatory and molecular characterization. CRE and ESBL-E isolates underwent species identification by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF Biotyper 3.1; Bruker Daltronics, https://www.bruker.com). We conducted antimicrobial susceptibility testing of CRE isolates by using reference broth microdilution with a metallo-β-lactamase screen (22,23), screening for carbapenemases by using the modified carbapenem inactivation method (24), and real-time PCR testing for bla_KPC, bla_NDM, bla_VIM, bla_IMP, and bla_OXA-48-like genes (25–28). If a CRE isolate harbored a carbapenemase-producing gene according to PCR, it was classified as carbapenemase producing (CP). We used real-time PCR to screen all isolates with a colistin MIC >2 µg/mL for plasmid-mediated colistin resistance genes (mcr-1 and mcr-2) (29). We conducted antimicrobial susceptibility testing of ESBL-E by using reference broth microdilution and performed phenotypic screening for ESBL production with ceftazidime and cefotaxime alone and in combination with clavulanate (24). We conducted whole-genome sequencing on a subset of CRE isolates from 2016–2018 that were confirmed to be carbapenem resistant and on the subset of ESBL isolates received from CDC (Appendix).

We calculated crude incidence rates by using case counts and 2016–2020 US Census estimates of the surveillance area population <18 years of age. For incidence rates presented by region, demographics, or age group, denominators represent the pediatric population also stratified by that subgroup. Analysis was limited to case report forms completed as of November 29, 2022. We performed descriptive and comparative analyses for CRE and ESBL-E cases by using the χ² test or the Fisher exact test (where applicable) for categorical variables and the Wilcoxon rank sum test for continuous variables. We used SAS version 9.4 (SAS Institute Inc, https://www.sas.com) to conduct data analyses.

During 2016–2020, a total of 159 incident CRE cases were identified in 142 children across 10 EIP sites. Of the 159 cases, 83 (52.2%) isolates were E. cloacae complex, 50 (31.5%) E. coli, 17 (10.7%) K. pneumoniae, 5 (3.1%) K. aerogenes, and 4 (2.5%) K. oxytoca (Table 1). The number of CRE cases per EIP site ranged from 3 to 47. E. coli comprised half or most of the CRE cases in New Mexico (50.0%), Tennessee (71.4%), and California (85.7%), whereas E. cloacae complex were the most common organisms at the other sites (44.4%–71.4%). Of the 142 unique persons with CRE, during the 5-year surveillance period >2 incident cultures were obtained from 17 (12.0%) (range 2–6 episodes).

During 2019–2020, a total of 207 incident ESBL-E cases were identified in 184 children across the 6 participating EIP sites. Of the 207 cases, 182 (87.9%) isolates were E. coli, 23 (11.1%) K. pneumoniae, and 2 (1.0%) K. oxytoca (Table 1). The number of ESBL-E cases per EIP site ranged from 14 to 57 cases; at all 6 sites, the predominant organism was E. coli (82.5%–100.0%). Of the 184 unique persons with ESBL-E, during the 1.5-year surveillance period, >2 incident cultures were obtained from 23 (12.5%) (range 2–4 episodes).

The overall annual CRE incidence rate (cases/100,000 pediatric population) across EIP sites during the 5-year period was 0.70 (range 0.47–0.87). The overall annual ESBL-E incidence rate during the 1.5-year period was 23.08, decreasing from 26.54 in 2019 to 19.63 in 2020. Crude incidence rates for CRE and ESBL-E varied by geographic region and year (Table 2).

During 2016–2020, annual incidence rates for infants (children <1 year of age) were consistently higher than those for other age groups, ranging from 1.95 to 3.82 cases/100,000 pediatric population for CRE and 46.85 to 91.97 cases/100,000 pediatric population for ESBL-E (Figure). In addition, crude annual incidence rates for CRE and ESBL-E were nearly always higher for female than male children (Table 2), except in the youngest age group. During 2016–2020, average annual crude incidence rates for CRE cases were higher for male than for female children <1 year of age (3.38 vs. 2.35 cases/100,000 pediatric population).

Of 159 CRE cases, 94 (59.1%) were in girls (Table 3), compared with 49.1% of the overall pediatric population. Median age was 5 (interquartile range [IQR] 1–10) years; 4 (2.5%) children were <1 month of age and 31 (19.5%) were 1–12 months of age (compared with 5.4% of the population <1 year of age). Most children with CRE were White (79 [50.0%]) and non-Hispanic (86 [54.1%]). Similarly, of 207 ESBL-E cases, most were in girls (165 [79.7%] compared with 49.1% of the overall pediatric population), White (92 [44.4%]), and non-Hispanic (96 [46.4%]). The median age was 6 (IQR 2–15) years; 3 (1.5%) children were <1 month of age, and 27 (13.0%) were 1–12 months of age (compared with 5.1% of the population <1 year of age).

Clinical characteristics were available for 158 CRE and 169 ESBL-E cases with completed case report forms (Table 3). Of those, a greater proportion of children with CRE than ESBL-E had a history of premature birth (20 [12.7%] vs. 11 [6.5%] among those with term birth; p = 0.06) and any underlying condition (99 [62.7%] vs. 59 [34.9%]; p<0.01).

For most CRE and ESBL-E cases, including those in children <1 year of age, organisms were isolated from urine (131 [82.4%] from CRE cases and 196 [94.7%] from ESBL-E cases) (Table 4). Accordingly, the most common reported infection type was lower urinary tract infection (89 [56.3%] for CRE and 125 [74.0%] for ESBL-E). For CRE and ESBL-E cases, the greatest number of isolates were collected in an emergency department or outpatient setting (108 [68.4%] for CRE and 154 [91.1%] for ESBL-E), although CRE cases were more likely than ESBL-E cases to be hospital onset (40 [25.3%] vs. 13 [7.7%]; p<0.01) (Table 3). Healthcare-associated community onset (68 [43.0%] CRE vs. 40 [23.7%] ESBL-E) and community-associated (43 [27.2%] CRE vs. 109 [64.5%] ESBL-E) represented most CRE and ESBL-E cases.

Among cases with available case report form data, a greater proportion of children with CRE than ESBL-E underwent acute care hospitalization (74 [46.8%] CRE vs 38 [22.5%] ESBL-E; p<0.01) or surgery (61 [38.6%] CRE vs. 16 [9.5%] ESBL-E; p<0.01) within 1 year before specimen collection (Table 4). In the 2 days before specimen collection, CRE cases were also more likely than ESBL-E cases to have a central venous catheter (37 [23.4%] CRE vs. 12 [7.1%] ESBL-E; p<0.01) or other indwelling device (excluding urinary catheter) (60 [38.0%] CRE vs. 19 [11.2%] ESBL-E; p<0.01). ESBL-E cases were significantly more likely than CRE cases to have no reported healthcare exposures (110 [65.1%] vs. 43 [27.2%]; p<0.01). Among ESBL-E cases, antimicrobial use was documented in the 30 days before date of incident specimen collection for 52 (31.8%).

Hospitalization at the time of or within 30 days of specimen collection was required for a greater proportion of community-onset CRE (40 [36.0%]) versus ESBL-E (13 [8.7%]) cases (p<0.01). Median duration of admission among all hospitalized community-onset and hospital-onset cases was 18 days (IQR 3–103 days) for those with CRE versus 10 days (IQR 4–43 days) for those with ESBL-E (p = 0.34).

The antimicrobial resistance profiles of incident CRE and ESBL-E cases from local clinical laboratories are shown elsewhere (Appendix Table 1). Among the 86 CRE isolates submitted for carbapenemase testing, 9 (10.5%) isolates from 6 of the 10 EIP sites harbored a carbapenemase: 4 bla_NDM, 3 bla_OXA-48-like, and 2 bla_KPC (Table 1). Distribution of CP-CRE varied by organism. The 9 CP-CRE isolates were from 9 children, fewer than half of whom were 1–3 years of age (3 [15.8%]) and 4–9 years of age (3 [13.0%]); 6 (18.8%) were from children with no reported underlying conditions (Appendix Table 2). The most common source was urine (8 CP-CRE isolates, 11.6% of submitted). Among the 11 CRE isolates from 2016–2018 that were sequenced, identified multilocus sequence types (STs) were diverse (Appendix Table 3); 2 of 11 isolates harbored CP genes. Separately, of the 7 ESBL-E organisms that underwent whole-genome sequencing at CDC, 6 were E. coli; ST131 and potential acquired ESBL gene bla_CTX-M-15 were most common (Appendix Table 4).