Limited detection of carbapenem-resistant Enterobacterales in veterinary patients at a Canadian referral hospital

Abstract

Objective

To estimate the prevalence of carbapenem-resistant Enterobacterales (CRE) fecal carriage in equine, canine, and feline patients at the Atlantic Veterinary College (Charlottetown, Prince Edward Island), a Canadian teaching and referral hospital.

Animals and procedure

Fifty-seven fresh, naturally voided fecal samples from 32 canine, 21 equine, and 4 feline Atlantic Veterinary College patients were collected and processed in June and July 2025. Samples were processed using selective culture, followed by MALDI-TOF MS and phenotypic assays for screening and detection of CRE and Enterobacterales species with reduced susceptibility to carbapenems. Relevant patient data were obtained to assess for associations with CRE and carbapenem susceptibility status.

Results

One canine Morganella morganii isolate was classified as CRE (sample-level prevalence: 1.8%), but carbapenemases were not detected. Twelve (21%) samples grew Enterobacterales on MacConkey plates supplemented with meropenem, indicating reduced susceptibility to carbapenems.

Conclusion and clinical relevance

The low prevalence of CRE within this Canadian sample is reassuring; however, the presence of CRE and high proportion with reduced susceptibility to meropenem underscore the importance of readying measures for early detection and control of CRE in veterinary patients and ensuring strong compliance with clinic infection-control practices. Carbapenem-resistant Enterobacterales outbreaks at Canadian veterinary hospitals, as documented elsewhere, remain a concern.

It is well established that antimicrobial resistance is an increasingly global phenomenon (1–5). Carbapenem resistance within the bacterial order Enterobacterales (EB), which includes Escherichia coli and Klebsiella pneumoniae, has been detected with increasing frequency in humans and animals (2,4–6). These organisms are known as carbapenem-resistant Enterobacterales (CRE). Resistance to carbapenems can be mediated by several mechanisms broadly classified as carbapenemase-producing CRE (known as carbapenemase-producing Enterobacterales, CPE) or non-carbapenemase-producing CRE. Carbapenemase genes are on plasmids that often carry resistance genes for multiple antimicrobial classes (4,6–9). The transferability of multiple antimicrobial resistance genes to competent bacterial strains renders plasmid acquisition of highest importance among carbapenem resistance mechanisms. Carbapenem use is often restricted in veterinary medicine to preserve the efficacy in humans (3,4).

Carbapenem-resistant Enterobacterales can cause a variety of opportunistic and healthcare-associated infections in animals, such as surgical- and catheter-site infections, pneumonia, and urinary tract infections (6). In addition, CRE can subclinically colonize the gastrointestinal tract of animals (6,9,10). Subclinical carriage of CRE may complicate early detection and response in a passive surveillance system that targets only clinical cases. Active surveillance may be needed to accurately estimate the prevalence of CRE in a population and prevent its spread in an outbreak. The use of selective enrichment media in bacterial culture can be used to increase the detection of CRE and other multidrug-resistant organisms (11,12). Because CRE can be transmitted between animals and humans, there are farreaching consequences of CRE in companion animals and transmission within a veterinary clinic environment. No CRE outbreaks have yet been reported in veterinary hospitals in Canada, but they have been increasingly identified in veterinary hospitals in the United States and elsewhere in the world, with important, documented effects on patients, personnel, and facilities (2–6,9).

To date, there has been minimal active surveillance conducted for CRE within Canadian veterinary patients (13). This study aimed to determine the prevalence of CRE fecal carriage in canine, feline, and equine cases admitted to the University of Prince Edward Island, Atlantic Veterinary College Veterinary Teaching Hospital (AVC VTH) (Charlottetown) during the summer of 2025.

Through convenience sampling, fresh, naturally voided fecal samples were collected from canine, feline, and equine inpatients and outpatients at the AVC VTH from June 12, 2025 to July 30, 2025. The AVC VTH personnel were instructed to collect ~5 g of feces as soon as possible after admission, and to submit all samples, regardless of the amount collected. Animals with confirmed or suspected infectious gastrointestinal disease or housed under isolation precautions were excluded from the study, to comply with facility infection control protocols. All other canine, feline, and equine inpatients and outpatients that naturally voided feces during their hospital stay were eligible for study inclusion, and AVC VTH personnel were requested to collect all such samples. Samples were kept cold in a cooler with ice packs or a refrigerator until analysis and for no longer than 24 h. This study was approved by the University of Prince Edward Island Animal Care Committee and Institutional Biosafety Committee.

Each fecal sample was plated onto MacConkey agar (Mac) using a sterile swab, to demonstrate the presence of baseline EB gastrointestinal microbiota of the animal. From the same sample, ~1 g of feces was added to 9 mL of lysogeny broth supplemented with 1 mg/L cefotaxime (LB+) (14). Before plating or adding to LB+, the sample was mixed by passing the swab through multiple areas of the fecal specimen. The Mac plate and LB+ were incubated overnight at 35°C ± 2°C. After incubation, 10 μL of LB+ was inoculated onto a Mac plate supplemented with 0.125 mg/L meropenem (Mac+M), to select for EB with reduced susceptibility to carbapenems (12). The Mac+M plates were incubated overnight at 35°C ± 2°C. After 24 h and 48 h incubation, Mac and Mac+M plates were evaluated for bacterial growth typical of EB, and up to 3 colony types per plate were identified using a MALDI Biotyper (Bruker Daltonik, Bremen, Germany), following the manufacturer’s instructions. Any EB identified on Mac+M were subcultured to Columbia agar supplemented with 5% sheep blood, for further characterization and storage.

Disk and gradient strip diffusion testing was used to determine the susceptibility of the EB recovered from Mac+M to meropenem, following Clinical and Laboratory Standards Institute (CLSI) methodology and interpretative criteria (15). The modified carbapenem inactivation method (mCIM) was used to determine if EB recovered from Mac+M were CPE, as described in the CLSI standard (15). All isolates were frozen at −80°C. Confirmed CRE were sent to the University of Guelph Animal Health Laboratory (Ontario) for whole-genome sequencing, following previously described protocols for short-read, paired-end sequencing (16). Genes conferring antimicrobial resistance were screened within the assembled genomes using the Comprehensive Antibiotic Resistance Database (CARD) database (17).

Each batch of selective culture agar and broth was tested for the growth of well-characterized CPE reported elsewhere (9,10) and for the suppression of E. coli ATCC 25922 (beta-lactam-susceptible) and K. pneumoniae ATCC 700603 (an extended-spectrum beta-lactamase producer). Antimicrobial susceptibility testing and mCIM quality control were carried out as described (15).

For each study animal, relevant patient demographic information was obtained from the electronic medical record system and included species, province of residence, primary reason for hospital visit, date of hospital admission, and reported antimicrobial administration within the previous 30 d. Antimicrobial use was further categorized by class of antimicrobials reportedly received and for beta-lactams, further classified by subgroup [penicillins, potentiated penicillins, cephalosporins (by generation), carbapenems]. Prevalence of CRE and CPE was defined as the proportion of unique patient samples with CRE and CPE over the number of samples tested. Exact logistic regression was used to identify associations between EB growth on Mac+M and patient species, province of residence, and reported antimicrobial administration within the previous 30 d. Confidence intervals (95% CI) for proportions (exact Clopper-Pearson) and odds ratios (ORs) were calculated. Statistical significance level was set at 5%. Data were managed using Excel software (Microsoft, Redmond, Washington, USA) and statistical testing undertaken using StataSE version 19 (StataCorp, College Station, Texas, USA).

A total of 57 samples from unique patients (1 sample per animal) were collected over the study period: 32 canine samples, 21 equine samples, and 4 feline samples. Electronic medical record data were available for 56 (98.2%) of the patients. Samples came from animals residing predominantly in Atlantic Canada (Table 1). Time between admission and sample collection ranged from 0 to 11 d (median: 1 d). The reason for the visit/presenting medical complaint was categorized based on review of the medical record, with some animals placed into > 1 category. Most animals were presented because of illness (51/56; 91.1%), with a minority (5; 8.9%) considered healthy (e.g., mares accompanying foals, blood-donor screening). Ill animals were most commonly presented for musculoskeletal-related concerns (n = 10), gastrointestinal signs (n = 9), surgical procedures (n = 7), and suspected infections (n = 6). Half of the samples were from animals that had reportedly received antimicrobials within the 30 d before sample collection (Table 1). Sixteen of the samples came from animals that had received a single class of antimicrobials; 12 samples were from animals that had received ≥ 2 classes. Of the 28 samples with previous antimicrobial exposure, beta-lactam exposure was most common (93.0%), followed by exposure to fluoroquinolones (25.0%), nitroimidazoles (14.0%), potentiated sulfonamides (11.0%), tetracyclines (7.0%), and aminoglycosides and macrolides (each 4.0%). No patient was reported to have received a carbapenem. Twelve of the 57 samples (21.1%; 95% CI: 11.4 to 33.9%) grew EB on Mac+M (Table 1). From these 12 samples, 16 EB species were isolated. Six E. coli isolates and 1 Enterobacter sp. isolate were recovered from 6 equine samples. Two E. coli isolates, 3 Morganella morganii isolates, and 2 Proteus mirabilis isolates were recovered from 5 canine samples. One E. coli isolate and 1 M. morganii isolate were isolated from 1 feline sample. One canine M. morganii isolate was resistant to meropenem (minimum inhibitory concentration > 32 μg/mL; susceptibility breakpoint ≤ 1 μg/mL). The remainder of EB isolates were susceptible to meropenem.

TABLE 1.

Demographic data for 56 veterinary patients from which fecal samples from 12 yielded Enterobacterales (EB) with reduced susceptibility to carbapenems (Atlantic Veterinary College, Charlottetown, Prince Edward Island, 2025). Enterobacterales isolated from MacConkey agar with 0.125 mg/L meropenem (Mac+M) were considered to have reduced susceptibility to carbapenems. Associations between patient factors and detection of EB with reduced susceptibility to carbapenems are reported (exact logistic regression).

Factor	Number (%)of samples	Number (%) with EB growth on Mac+M	OR (95% CI)	P-value
Species				0.6
Canine	32 (56.1)	5 (15.6)	Referent
Equine	21 (36.8)	6 (28.6)	2.1 (0.5 to 10.5)	0.4
Feline	4 (7.0)	1 (25.0) 1.8	(0.03 to 28.1)	1.0
Province of residence				0.2
Prince Edward Island	25 (44.6)	2 (8.0)	Referent
New Brunswick	12 (21.4)	3 (25.0)	3.7 (0.4 to 51.0)	0.4
Newfoundland and Labrador	1 (1.8)	0 (0.0)	12a (0 to 468)	1.0
Nova Scotia	16 (28.6)	6 (41.2)	6.5 (1.0 to 77.3)	0.06
Ontario	2 (3.6)	1 (50.0)	9.6 (0.1 to 939.8)	0.4
Antimicrobial in previous 30 d
No	28 (50.0)	2 (7.1)	Referent
Yes	28 (50.0)	10 (35.7)	7.0 (1.3 to 73.0)	0.02
Beta-lactam in previous 30 d				0.1
Cephalosporins (1st generation)	10 (38.5)	2 (20.0)	Referent
Cephalosporins (3rd generation)	2 (7.7)	0 (0.0)	2.0a (0 to 33.1)	1.0
Penicillins	7 (26.9)	5 (71.4)	8.4 (0.7 to 162.7)	0.1
Potentiated penicillins	7 (26.9)	2 (28.6)	1.6 (0.09 to 28.3)	1.0

CI — Confidence interval; OR — Odds ratio.

^aMedian unbiased estimate.

All the EB isolated from Mac+M were negative for carbapenemase production, as determined by mCIM. Patient-level prevalence of CRE was 1.8% (95% CI: 0.04 to 9.4%). Patient-level prevalence of CPE was 0% (95% CI:0.0 to 6.3%). Whole-genome sequencing confirmed the meropenem-resistant M. morganii isolate did not contain a carbapenemase gene. A homologue of the kpnH gene, which encodes a major facilitator superfamily antibiotic efflux pump that confers decreased susceptibility to several antimicrobial classes, including carbapenems, was detected (78.9% nucleotide match with 99.6% coverage). Resistance genes for amphenicols, fluoroquinolones, and tetracyclines were also detected with > 90% nucleotide match and coverage (NCBI project PRJNA325243; BioSample SAMN52930523). Recent antimicrobial exposure was positively associated with EB growth on Mac+M (OR: 7.0; 95% CI: 1.3 to 73.0) (Table 1). Patient species, province of residence, and subgroup of beta-lactam reportedly received within the past 30 d were not significantly associated with EB growth on Mac+M (Table 1).

In the present study, we report the limited detection of CRE fecal carriage in Canadian veterinary patients. The isolation of 1 non-carbapenemase-producing carbapenem-resistant M. morganii from this population suggests a currently low presence of CRE and a low-to-absent prevalence of CPE in companion animals in Atlantic Canada. This is consistent with a low reported prevalence of CRE in the eastern Canadian human population (18,19).

To date, there have been no reported outbreaks of CRE and, before our work, no reported detection of CRE in patient samples or in patients at Canadian veterinary teaching hospitals. Although this is promising, there has been minimal active surveillance for CRE presence in this population. One study at a western Canadian veterinary teaching hospital did not detect CRE in 60 equine patients; however, that study did not select for carbapenem-resistant isolates using supplemented culture media, as was done in the present study (13). As recent global trends in veterinary patients emphasize, there is an expected continued increase of antimicrobial resistant organisms, including CRE, with Canada likely following suit (2,4–6). Our findings, as well as the inevitable increase in CRE across the region, highlight the need for ongoing active veterinary surveillance for these organisms in the region, and for the development and rapid deployment of targeted surveillance and outbreak mitigation strategies for veterinary clinical practices.

Equally important as the identification and control of CRE spread in veterinary facilities are factors associated with (and ideally causal to) CRE and CPE emergence. In the present study, recent antimicrobial use was associated with an increase in the odds of recovering an EB species with reduced susceptibility to carbapenems. This is perhaps unsurprising, as previous work has shown recent antimicrobial use to be a risk factor for the recovery of CRE from animals (3). Additional work is needed to determine the specific antimicrobial drivers for this resistance (e.g., drug class); however, antimicrobial stewardship is clearly an important component in reducing CRE and CPE in animals.

Given the small sample size of the current study, further testing is needed to confirm the findings. However, based on the demographics and medical histories of patients sampled, it appears we included an appropriate representation of animals with known risk factors for CRE, including recent antimicrobial exposure (3,4,6). The necessity of relying on reported patient medical data in the electronic medical record was a limitation of this study, as this approach relies on owners recalling past information and questioning and documentation by veterinary personnel. As such, patient demographics and history may not be accurately represented.

This work serves as an important foundation for future CRE/CPE surveillance and research in Canadian veterinary populations. Veterinary surveillance programs for CRE should not focus solely on the detection of clinical CRE isolates; inclusion of methods to detect CRE fecal carriage is strongly encouraged.

Funding Statement

The authors also thank the US FDA Veterinary Laboratory Investigation and Response Network (Vet-LIRN) and Dr. Durda Slavic and Dr. Aparna Krishnamurthy at the University of Guelph for their financial and technical support with the whole-genome sequencing.