β-Lactam and Fluoroquinolone-Resistant Enterobacteriaceae Recovered from the Environment of Human and Veterinary Tertiary Care Hospitals

Introduction

The dissemination of antimicrobial-resistant bacteria in healthcare environments represents an important nosocomial threat in both veterinary and human hospitals. Bacteria resistant to extended-spectrum cephalosporin and carbapenem antimicrobials are especially problematic because their resistance genes are frequently located on highly mobile genetic elements, facilitating horizontal transmission. The extended-spectrum cephalosporin and carbapenem-resistant Enterobacteriaceae have been identified by the Centers for Disease Control and Prevention as serious and urgent threats to the public health, respectively (CDC 2013). The most common mechanism of bacterial resistance to these drugs is enzymatic through the production of β-lactamases or carbapenemases. β-lactamase resistance genes of clinical importance in both veterinary and human medicine include the AmpC bla _CMY and the extended-spectrum β-lactamase (ESBL) bla _CTX-M. Carbapenemase resistance genes, primarily bla _KPC, are widely disseminated in U.S. healthcare institutions but have only rarely been reported in animals in the United States.

Fluoroquinolones and quinolones are another important group of broad-spectrum antimicrobials commonly used to treat local and systemic infections in humans and animals, among which ciprofloxacin is one of the top three drugs currently used in medicine (Redgrave et al. 2014). Many bacterial mechanisms of fluoroquinolone resistance have been described, mainly chromosomal mutations resulting in target alterations, increased efflux, and reduced membrane permeability. As with other antimicrobials, fluoroquinolone resistance has been increasing and it is proportionally associated with their level of use and/or consumption (Redgrave et al. 2014).

Healthcare settings are thought to play a critical role in the emergence and amplification of antimicrobial-resistant bacteria due to the chronic selection pressure resulting from the frequent application of antibiotics (Otter et al. 2011). Among stool samples obtained from hospitalized patients that were submitted for Clostridium difficile screening, 10%, 3%, and 0.3% were reported with resistant Enterobacteriaceae harboring bla _CMY, bla _CTX-M, and bla _KPC, respectively (Mollenkopf et al. 2015). However, not only do hospitalized patients harbor these clinically important resistant bacteria, but also hospital surfaces have been implicated in the spread of antimicrobial-resistant bacteria. For example, hospital rooms with a prior occupant with an antimicrobial-resistant bacteria are a risk factor for patient acquisition of antimicrobial-resistant bacteria, suggesting that surfaces can play a role in nosocomial transmission even without direct contact between patients (Huang et al. 2006, Nseir et al. 2011). In addition, hospital disinfection can reduce the spread of nosocomial antimicrobial-resistant bacteria, providing further evidence of the important role of the hospital surfaces as a potential reservoir (Hayden et al. 2006).

In veterinary medicine, companion animals have been reported to have clinical infections with AmpC, ESBL, and carbapenemase-producing Enterobacteriaceae (CPE) bacteria (O'Keefe et al. 2010, Shaheen et al. 2011, 2013). Veterinary hospital environments may also play a role in the transmission of antimicrobial-resistant bacteria. Hospitalization is a risk factor for ESBL colonization in dogs, with ∼8% of companion animal patients reported to acquire a multidrug-resistant Escherichia coli during hospitalization (Gibson et al. 2011, Hamilton et al. 2013). Pets and their owners are frequently in close contact in shared living space, which can facilitate the spread of antimicrobial-resistant bacteria between humans and pets (Song et al. 2013). In addition, pet owners are at increased risk of colonization with extended-spectrum cephalosporin-resistant E. coli, indicating that pet ownership plays a role in transmission of antimicrobial-resistant bacteria (Meyer et al. 2012).

Our objective is to determine the relative frequency of environmental contamination of large metropolitan university-associated human and veterinary hospitals with bacteria expressing AmpC, ESBL, carbapenem-resistant, and fluoroquinolone-resistant phenotypes.

Methods

Two large university human teaching hospitals in Columbus, Ohio, The Ohio State University Wexner Medical Center's University Hospital and University Hospital East, were each sampled five times, between July 2014 and December 2014. Electrostatic cloths (Swiffer Dusters; Proctor and Gamble, Cincinnati, OH) were used to obtain duplicate composite samples of surfaces from surgery, internal medicine, intensive care, and oncology service areas of the hospital, and common areas, such as elevators and patient transport vehicles (Supplementary Table S1; Supplementary Data available online at www.libertpub.com/vbz). In addition, The Ohio State University Veterinary Medical Center (VMC), a large veterinary referral hospital located in the same metropolitan area, was sampled three times between June 2014 and January 2015. Electrostatic cloths were employed to obtain a composite sample of surfaces from internal medicine, intensive care, oncology, dermatology, and surgery service areas, and common areas of the companion animal section of the hospital (Supplementary Table S2). In addition, samples were acquired from the farm animal section of the hospital (Supplementary Table S2). All electrostatic cloth samples from the human hospitals and the VMC were collected by the same team of two trained individuals.

Electrostatic cloths are immersed in 100 mL of nutrient broth modified with 2 μg/mL of cefotaxime, incubated overnight at 37°C, and the broth inoculated onto three MacConkey agar plates modified with 8 μg/mL of cefoxitin, 4 μg/mL of cefepime, and 1 μg/mL meropenem to screen for the AmpC, ESBL, and CPE phenotypes, respectively. Bacterial species were identified using MALTI-TOF, and β-lactamase genotypes of isolates were confirmed using previously reported primers for bla _CMY (Winokur et al. 2001, Kanwar et al. 2014) and bla _CTX-M (Wittum et al. 2010), and CPE genotypes of isolates were confirmed using bla _KPC primers (Peirano et al. 2011).

Parallel screening for fluoroquinolone resistance was accomplished by immersing the duplicate electrostatic cloth in 100 mL of nutrient broth with 16 μg/mL of nalidixic acid followed by inoculation to two MacConkey agars with 16 μg/mL nalidixic acid and 2 μg/mL ciprofloxacin after overnight incubation. New unused electrostatic cloths taken directly from the same packaging as those used for sampling were similarly processed for quality control purposes.

Results

For hospital 1, 15 surfaces from surgery, intensive care, internal medicine, and oncology regions and 2 surfaces from shared equipment were sampled during each of 5 visits, resulting in 85 total surfaces. E. coli harboring bla _CMY were recovered once (1.2%, 95% CI: 0–3.5%) from a patient transport vehicle, but no isolates with bla _CTX-M were detected (Table 1). A single Klebsiella pneumoniae (1.2%, 95% CI: 0–3.4%) harboring bla _KPC was recovered from the intensive care waiting room. Only one isolate (1.2%, 95% CI: 0–3.5%) recovered from a medical trolley was found resistant to fluoroquinolones. In hospital 2, 12 surfaces from surgery, intensive care, and internal medicine regions and 2 shared equipment surfaces were sampled during each of five visits, resulting in 70 total surface samples. No isolates harboring bla _CMY were recovered, whereas one E. coli (1.4%, 95% CI: 0–4.2%) harboring bla _CTX-M and one K. pneumoniae (1.4%, 95% CI: 0–4.2%) harboring bla _KPC were recovered from a trolley in the surgery service (Table 1).

None of the isolates obtained in hospital 2 showed resistance to fluoroquinolones. In the VMC, 34 surfaces were sampled repeatedly from general areas within the internal medicine, intensive care, oncology, dermatology, and surgery services, and in common areas of the companion and farm animal sections of the hospital over the course of three visits. E. coli harboring bla _CMY was recovered from 100% of the surfaces (n = 20) sampled in the farm animal section, and from 18.2% of the surfaces (12/66, 95% CI: 8.9–27.5%) sampled in the small animal section of the VMC. E. coli harboring bla _CTX-M were recovered from 55% of surfaces (11/20, 95% CI: 33.2–76.8%) sampled in the farm animal section, and from 16.7% of surfaces (11/66, 95% CI: 7.7–25.7%) sampled in the small animal section. No CPE isolates were recovered from surfaces in the VMC. Resistance to fluoroquinolones was observed from 45.0% (9/20, 95% CI: 23.2–66.8%) of the surfaces sampled at the farm animal hospital, whereas 37.9% (25/66, 95% CI: 26.2–49.6%) of samples with fluoroquinolone resistance was observed at the companion animal hospital (Table 1).

Discussion

The veterinary hospital had a much higher proportion of surfaces contaminated with extended-spectrum cephalosporins (ESC) and fluoroquinolone-resistant bacteria than human hospital surfaces. Veterinary hospitals, by nature of their animal patients, are more likely to have fecal flora contamination of hospital environment surfaces. We found that all of the surfaces sampled in the farm animal section of the veterinary hospital were contaminated with Enterobacteriaceae expressing the AmpC phenotype, and 55% were contaminated with Enterobacteriaceae expressing the ESBL phenotype. Livestock have been reported to commonly shed E. coli harboring the AmpC bla _CMY and the ESBL bla _CTX-M. This may be the result of the frequent application of ceftiofur for the treatment and control of various infectious diseases in livestock populations. Our lower recovery of AmpC and ESBL phenotypes for the companion animal and human hospitals may reflect the lower frequency of ESC applied in these patient populations.

The companion animal hospital exhibited a higher proportion of surfaces with fluoroquinolone-resistant isolates than the farm animal and human hospitals. This result may be related to the frequency of use of fluoroquinolones for therapy in companion animals. Veterinarians are restricted from extralabel prescribing of fluoroquinolones in food animal species, but similar restrictions are not in place for companion animals. Our results highlight the significant role that companion animals can play as reservoirs for antimicrobial resistance. Resistant bacteria from pets can be spread to owners, cohabitant animals, and/or the hospital environment, indirectly exposing hospital personnel and other patients. Even though the prevalence of fluoroquinolone-resistant bacteria may vary between geographical regions, some clinical researchers have suggested that these drugs should be used only as second-line agents during treatments (Redgrave et al. 2014).

We recovered CPE from the environment of the human hospitals but not the veterinary hospital. These multidrug-resistant organisms have been classified as an “urgent” public health threat by the U.S. CDC because nosocomial infections result in a case fatality rate approaching 50% (CDC 2013). Thus, their presence in the hospital environment represents an important nosocomial infection risk. Carbapenem drugs are an important therapy for invasive Gram negative infections in humans, but they are rarely used in companion animals and are restricted from use in food animals. This difference in carbapenem use between human and veterinary medicine likely impacted our recovery of CPE from hospital environments.

Hospital surfaces contaminated with antibiotic-resistant bacteria may pose a risk to patients. A study in France reported that 4% of surfaces within human hospital rooms housing children colonized or infected with ESBL-harboring Enterobacteriaceae were subsequently contaminated with ESBL-harboring Enterobacteriaceae (Weber and Rutala 2013). Although we did not sample patient rooms, they may be at higher risk of surface contamination resulting from patient and staff contact.

We found that areas in the human hospitals that would normally be expected to have a low risk were contaminated with both ESC-resistant bacteria and CPE, although at low frequency. These surfaces, including the intensive care waiting rooms, patient transport vehicle, and hospital carts, likely have limited direct contact with patients. They might, therefore, be perceived as low risk surfaces for bacterial contamination leading to relaxed precautionary behaviors by both patients and staff. Effective cleaning and disinfection protocols, together with individual infection control practices such as appropriate hand hygiene, are required in both human and veterinary hospitals to protect patients from antibiotic-resistant bacteria disseminated in hospital environments.

Our results suggest that hospital medical trolleys or carts that are designed to move easily between patient rooms can serve as fomites for the movement of clinically relevant resistant bacteria, including CPE. Others have reported the contamination of hospital carts with CPE and the potential role of carts in the nosocomial transmission of resistant pathogens (Chae et al. 2018). Carts used in veterinary hospitals have also been reported to be contaminated with clinically important resistant pathogens, including methicillin-resistant Staphylococcus aureus (van Balen et al. 2013). A better understanding of the role of hospital carts in dissemination, and the importance of effective cleaning and disinfection, may provide an effective intervention to reduce nosocomial transmission of antibiotic-resistant pathogens.

One limitation of this study is that the inherent differences in human and veterinary hospital environments and patient populations make interpretation of observed differences in the frequency of recovery of specific bacterial resistance genotypes difficult. These inherent differences resulted in differences in the surfaces that were sampled and in the intensity of sampling between the human and veterinary hospitals, which may have impacted our results. For example, floor drains likely to harbor bacterial contamination were sampled in the veterinary hospital due to their accessibility, but drains were not available for sampling in the common areas of the human hospitals. However, these results can be used to provide a better understanding of the risk of nosocomial transmission of antibiotic-resistant bacteria in diverse hospital environments.