Trends in resistant Enterobacteriaceae and Acinetobacter species in hospitalized patients in the United States: 2013–2017

Antibiotic-resistant Gram-negative bacteria have been recognized as a fundamental risk to patient health on both national and global levels [1‐3]. Enterobacteriaceae, which account for a significant proportion of infections in hospitalized patients in the US [4, 5], are of particular concern. Carbapenem-resistant Enterobacteriaceae (CRE) have limited treatment options and thus pose a significant clinical dilemma, but extended-spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae and multidrug-resistant (MDR) Enterobacteriaceae are also difficult to treat and occur at a higher prevalence, therefore endangering a greater number of patients [6]. The Centers for Disease Control and Prevention (CDC) categorizes CRE and ESBL-producing Enterobacteriaceae as urgent and serious threats, respectively, and the World Health Organization (WHO) considers them a critical priority for drug development [1‐3]. Carbapenem-resistant and MDR Acinetobacter are also featured on the CDC and WHO lists of dangerous pathogens [1‐3]. Although Acinetobacter infections are relatively rare, this pathogen is difficult to treat due to high rates of antibiotic resistance and paucity of options [7].

Trends in resistance patterns provide important insights into emerging pathogens as well as inform public health, infection control, and antimicrobial stewardship approaches. Although comprehensive data on hospital-acquired infections, including catheter-associated urinary tract infections (UTIs) and central-line associated bloodstream infections (BSIs), are available from the US National Healthcare Safety Network (NHSN) [6, 8], national comparative data for antimicrobial-resistant Gram-negative pathogens in the complete population of hospitalized patients are not as easily accessed. The objective of this study was to examine trends in resistance in selected Gram-negative pathogens collected from US hospitals from 2013 through 2017 based on rates of resistance per 100 hospital admissions and proportions of resistant isolates.

A total of 411 hospitals provided data for this study (Table 1). About three-quarters of hospitals (76.6%) were classified as urban, 70.6% were non-teaching hospitals, and the most common bed size was 100 to 300 (42.8%). Geographically, the largest concentration of hospitals was in US Department of Health & Human Services Region 4 (south central states; 23.6%) and Region 5 (north central states; 23.4%) (Table 1).

More than 1 million Enterobacteriaceae isolates were evaluated for ESBL (1,112,312 tested), Carb-NS (1,275,311 tested), and MDR (1,275,311 tested). Over the 5-year period, 12.05% Enterobacteriaceae isolates were identified as ESBL phenotype (hereafter referred to as ESBL Enterobacteriaceae), 1.21% as Carb-NS, and 7.08% as MDR (Table 2). A total of 19,325 Acinetobacter spp. isolates were tested for Carb-NS and MDR. Of these isolates, 37.48% isolates were identified as Carb-NS and 47.66% as MDR.

For Enterobacteriaceae, urine cultures accounted for the majority of resistant isolates (66.66% ESBL, 45.87% Carb-NS, and 63.04% MDR), followed by skin/wound (11.89, 20.53, and 13.12%, respectively) (Table 2). The highest rates of resistance were observed in respiratory cultures (17.41, 3.44, and 10.23%, and for ESBL, Carb-NS, and MDR, respectively). For Acinetobacter spp., respiratory cultures were the most common source of Carb-NS (39.94%) and MDR (39.29%), and also had the highest proportion of Carb-NS and MDR isolates (45.00 and 56.29%, respectively) (Table 2). Skin/wound was the second most common source of resistant Acinetobacter spp. (36.08% for Carb-NS, 36.04% for MDR).

Observational data for rates and proportions of antibiotic-resistant Enterobacteriaceae and Acinetobacter spp. from 2013 to 2017 are shown by quarter in Additional files 2 and 3, respectively.

Antimicrobial resistance in Gram-negative pathogens continues to threaten public health and increase the societal cost of health care. Carbapenem-resistant pathogens are a particular problem due to limited treatment options, ability to infect multiple organ systems, and high attributable mortality and cost [13, 20]. Although CRE are unquestionably an important clinical concern, ESBL and MDR Enterobacteriaceae pose a substantial threat to patient health and are present in greater numbers [1, 6]. Based on our Q4 2017 data, these pathogens accounted for a 10-fold higher incidence of pathogens in hospitalized patients than Carb-NS Enterobacteriaceae. Acinetobacter spp., a rare but dangerous Gram-negative pathogen, is associated with high rates of resistance and mortality [7, 21].

Information on trends in susceptibility is important from public health, infection prevention, and antimicrobial development standpoints. Some recent reports suggest that rates of antimicrobial-resistant Enterobacteriaceae are decreasing, including an NHSN study of central line-associated BSIs and catheter-associate UTIs [8]. However, other studies suggest a continued increase in the rate of antimicrobial-resistant Gram-negative pathogens [22‐24]. Outside of hospitals, there has been an increase in antibiotic-resistant Gram-negative pathogens in poultry and livestock [25], which may contribute to the spread of infections in the community.

Our study evaluated the rates of antimicrobial-resistant or nonsusceptible Enterobacteriaceae and Acinetobacter spp. from multiple culture sources in hospitalized patients throughout the United States from 2013 to 2017. We found that rates of ESBL and Carb-NS Enterobacteriaceae per 100 hospital admissions increased significantly during this time period, although increases were fairly modest, while rates of MDR Enterobacteriaceae, MDR Acinetobacter spp., and Carb-NS Acinetobacter spp. decreased. Analyses of the proportions of resistant isolates largely mirrored the trends observed with rates per 100 hospital admission, although increases in the proportion of Carb-NS Enterobacteriaceae and decreases in the proportion of Carb-NS Acinetobacter spp. did not reach statistical significance. For Acinetobacter spp., decreases were fairly constant throughout the 5-year period. In contrast, MDR Enterobacteriaceae rates and proportion of resistant isolates initially increased followed by a decrease from Q2 2015 to Q4 2017. Although this recent trend is encouraging, decreases in MDR Enterobacteriaceae were modest and rates are still high (6.8% in Q4 2017).

Independent reports of antimicrobial resistance and non-susceptibility are difficult to compare because they are influenced by a number of factors, including number of participating institutions, culture site(s), pathogens evaluated, definitions of resistance, culture ordering practices, and pathogen sources (community- versus hospital-acquired or surveillance versus infection). There are thus several reasons why our findings of increases in ESBL and Carb-NS Enterobacteriaceae might differ from studies reporting decreased rates of antimicrobial-resistant Enterobacteriaceae. Most notable is that our analyses include cultures collected in the admission period as well as those cultured in the hospital-onset time period; our analyses were confined to hospitalized patients, but were not specific to hospital-acquired infections. In contrast, CDC reports of hospital-acquired infections, by definition, attempt to exclude patients admitted with infections. In many facilities, hospital-acquired infections are targeted for intensive infection control and antimicrobial stewardship efforts, which may result in decreased antibiotic resistance in these infections compared with the resistance profile of pathogens in the community at large. The geographic representation of included hospitals can also influence results. We have shown previously that there are significant differences in rates of antibiotic-resistant Acinetobacter spp. and MDR Enterobacteriaceae across geographic regions in the US [12].

Our results concerning recent decreases in antibiotic-resistant Acinetobacter infections are consistent with other observations [6, 26]. Others have also observed seasonal variations in US Acinetobacter infections [27, 28], but previous studies have reported increased infection rates in summer months compared with winter months, whereas we found increased proportions of resistant pathogens (but not rates per 100 admissions) in the winter months. We observed a similar pattern for MDR Enterobacteriaceae, consistent with a report on antibiotic-resistant E. coli in which higher proportions of resistant isolates in winter months correlated with antibiotic prescribing practices [29]. The reasons for discrepancy in seasonal trends among different studies may involve differences in study methodology, in particular the inclusion of admission period cultures rather than solely “hospital onset” isolates and assessment of the proportion of resistant isolates versus overall frequencies of infection.

Limitations of our study include the collection and analysis of data from non-duplicated unique collected cultures rather than from unique patients. We are therefore unable to evaluate clinical outcomes associated with the antibiotic-resistant pathogens in this study. The results reported here represent non-duplicate culture positive isolates and not confirmed invasive infections. Our study included only selected Acinetobacter spp., and mechanisms of resistance were not investigated. Susceptibility was based on local microbiology practices at each facility and not standardized across facilities. Enterobacteriaceae ESBL and CRE testing practices and breakpoints are known to vary among different institutions [30], and institutions with lower breakpoints typically report higher resistance rates [31]. In particular, delayed implementation of updated Clinical & Laboratory Standards Institute guidelines for third-generation cephalosporins, cefepime, and carbapenem breakpoints may have influenced susceptibility assessments. These breakpoints were lowered in 2010 (carbapenems and third-generation cephalosporins) and 2014 (cefepime) for Enterbacteriaceae and in 2014 (carbapenems) for Acinetobacter spp., but immediate adoption in hospital laboratories likely varied due to the use of automated antimicrobial susceptibility testing systems. In one study of 25 US community hospitals, only 5 (20%) had adopted the 2010 carbapenem breakpoints by the end of 2012 [31]. It is possible that the increased resistance observed in ESBL and Carb-NS Enterobacteriaceae in our study may reflect the expanding adoption of these lower breakpoints, leading to the identification of a greater number of isolates as antibiotic resistant. The inclusion of isolates with intermediate resistance increased the number of isolates slightly, so our data of nonsusceptible isolates cannot be strictly compared to studies of pathogens meeting criteria for resistance. Antimicrobial susceptibility results for ertapenem may have also influenced Carb-NS rates, as AmpC hyperproduction and porin changes can cause some isolates to be Carb-NS to ertapenem while remaining susceptible to other carbapenems [32]. Despite these limitations, we believe our study provides an epidemiological lens on trends in antimicrobial resistance that reflects both the community and inpatient burden of hospitalized patients and may help inform further studies.

Our findings highlight the multiple factors influencing infections with resistant pathogens. Antibiotic-resistant Gram-negative bacteria are found in multiple culture sources, often spread from non-infected fomites and contacts, and can be ubiquitously represented in intensive care, medical, and surgical hospital wards [33]. Although understandably a focus of attention, reportable healthcare-acquired infections represent only a portion of clinically relevant infections. The inclusion of multiple sources of positive cultures may represent a more accurate risk of spreading drug-resistant organisms between patients, healthcare workers and family members [34, 35] or from community food supplies [25]. Additional Gram-negative bacteria, particularly Pseudomonas aeruginosa, also pose important clinical challenges. Work is in progress to explore trends in antibiotic-resistant P. aeruginosa.

Our findings from this large-scale study of patients in US hospitals show increasing numbers of infections due to ESBL and Carb-NS Enterobacteriaceae between 2013 and 2017. These data support continuing efforts by the CDC and WHO to combat these pathogens. Infections caused by MDR and Carb-NS Acinetobacter spp. are decreasing, but Acinetobacter remains a dangerous and difficult-to-treat pathogen. Continued infection control efforts, together with diagnostic and antimicrobial stewardship and new antibiotics to expand treatment options, will be required to manage these antibiotic-resistant Gram-negative pathogens.