Carbapenem-Resistant Organisms in Companion Animals in New York City, 2019–2022

Abstract

Carbapenem-resistant organisms (CROs) are a type of antibiotic-resistant bacteria that threaten human health. They can infect or colonize dogs and cats, with potential for zoonotic transmission to humans, but their prevalence in pet populations is not well described. To characterize CRO prevalence among gram-negative cultured isolates from New York City dogs and cats, we analyzed antimicrobial susceptibility data from a commercial veterinary diagnostic laboratory serving New York City veterinarians during 2019–2022. Among 16 115 gram-negative isolates, 256 (1.6%) were CROs cultured from 241 dogs and cats. CRO detections and the percentage positivity fluctuated during 2019–2022 and differed across the city's 5 boroughs. Data sharing between public health and veterinary diagnostic laboratories can identify CROs in pets and create opportunities to improve veterinary outreach and control of CROs in companion animals.

Graphical Abstract

Graphical Abstract.

This graphical abstract is also available at Tidbit: https://tidbitapp.io/tidbits/carbapenem-resistant-organisms-in-companion-animals-in-new-york-city-2019-2022-4b9ac4f4-e9c1-4263-bb3b-77f6b5586274?utm_campaign=tidbitlinkshare&utm_source=ITP

The emergence and spread of antimicrobial-resistant bacteria are pressing public health challenges affecting both human and animal health. Antibiotics are essential for treating infectious bacterial diseases, and when antibiotics are no longer effective bacterial infections can become difficult to treat. Antimicrobial-resistant bacteria were associated with approximately 4.95 million human deaths globally in 2019 [1]. While many animal species can be infected or colonized by resistant bacteria, much remains unknown about the global impact of resistant bacteria on animal health and the role of animals in community spread [2–6]. Since antimicrobial resistance is a multifaceted challenge, collaborations and data sharing between human and animal health sectors are necessary to understand and control antimicrobial resistance in humans and animals.

Carbapenem-resistant organisms (CROs) are an emerging type of resistant bacteria that are concerning to public health. CROs are bacteria resistant to carbapenem antibiotics, such as meropenem, imipenem, and ertapenem. Gram-negative CROs are of particular public health concern because carbapenem antibiotics are used as drugs of last resort for gram-negative antibiotic-resistant bacterial infections in people [7]. Carbapenems are not approved for use in veterinary medicine, although they are used occasionally in companion animals in an extra-label fashion [8].

Carbapenem resistance can be caused by several different mechanisms [9]. Carbapenemases, a type of enzyme that inactivates carbapenems, are the most concerning resistance mechanism because genes encoding carbapenemases are frequently carried on mobile genetic elements (eg, plasmids), facilitating rapid spread within and between bacterial populations [10].

While CROs have been detected in companion animals, public health entities are seldom notified about these detections and therefore are unable to pursue phenotypic testing for carbapenemase production or molecular identification of carbapenemase genes that are needed to guide public health action [11]. Without notification and additional tests for CROs, public health entities are limited in their ability to support veterinary professionals with CRO infection prevention and control efforts.

New York City (NYC) has historically been an epicenter for emerging resistant organisms [12, 13]. In a collaboration between the New York City Department of Health and Mental Hygiene and a large commercial veterinary diagnostic laboratory, we aimed to characterize CRO prevalence among gram-negative isolates from NYC dogs and cats. We retrospectively analyzed companion animal (dog and cat) antimicrobial susceptibility testing (AST) data for specimens submitted by NYC veterinarians during a 4-year period. Our study focused on gram-negative CROs because the World Health Organization and other public health agencies have deemed them priority resistant organisms. In addition, companion animals live in close proximity to their human guardians and caregivers, enabling ample opportunity for zoonotic transmission [7, 14, 15].

METHODS

We evaluated all bacterial isolates tested at a single commercial veterinary diagnostic laboratory for culture and antimicrobial susceptibility from dog and cat specimens submitted by veterinarians in NYC from 1 January 2019 through 31 December 2022. Specimens were submitted from veterinary clinics, animal care shelters, and veterinary referral hospitals. Each isolate was associated with a unique deidentified animal patient number, patient species, patient age, year of sample submission, 5-digit zip code of submitting facility, the source where the sample was obtained from the patient (eg, general specimen source information, such as nasal, skin, etc), the site where the sample was taken from the patient (eg, more specific source information, such as nasal flush, bite wound, etc), organism identified, and the organism's AST result for the carbapenems for tested at this laboratory, imipenem and meropenem. Antimicrobial susceptibility status was reported as “susceptible,” “resistant,” or “intermediate.” As there are no veterinary specific breakpoints to determine carbapenem susceptibility, breakpoints from humans are used in veterinary microbiology [16]. Susceptibility testing was performed using techniques including the Vitek (bioMérieux) automated system [17] and Kirby-Bauer testing [18]. This study did not include factors necessitating consent.

Data were cleaned, deduplicated, and analyzed using SAS Enterprise Guide 8.3 (SAS Institute). We conducted descriptive analyses including epidemiologic characteristics and temporal patterns of CROs in dogs and cats. Fungal isolates and isolates with an environmental source were removed from the dataset. Many animal patients had multiple or repeated cultures, so to avoid overestimating CRO prevalence, we deduplicated isolates. Isolates were considered duplicates if they shared the same unique patient number and same organism by genus (Figure 1). Multiple isolates were attributed to a single patient if a different genus and species were detected. Initial CRO detections, identified as isolates with an earlier collection year were retained during deduplication.

Figure 1.

Data deduplication of isolates from animal care facilities in New York City identified by culture and antimicrobial susceptibility testing by a single commercial diagnostic laboratory, 2019–2022.

A CRO isolate was defined as a gram-negative bacterial isolate with AST results indicating resistance to either meropenem and/or imipenem. We restricted analysis to gram-negative bacteria tested for carbapenem susceptibility [16]. All organisms were categorized and described by genus and species (Supplementary Table 1).

Age data were categorized in 5 groups, based on age groups established for cats: 0–4 (youth), 5–9 (early midlife), 10–11 (late midlife), 12–13 (senior), and 14–25 years (geriatric) [19]. Dogs and cats with missing ages and those aged >25 years were categorized as age unknown; the latter were likely data errors and not reflecting a true age. We used cat age groups for both species because these life stages are appropriate for cats and small and medium-breed dogs (just not large-breed dogs). Breed and size, which determine canine life stages, were unavailable.

The variable “composite site” combined data from both source and site. The composite site comprised 12 nonsterile and 9 normally sterile categories, as well as other and unknown sites (Supplementary Table 2) [20]. Urine samples were not considered sterile [21, 22]. If the source was “other,” the site variable was used to define the composite site, if possible. The composite site category of “other” included sources that were too vague to allow for categorization, and “unknown” included samples with a missing or unknown source.

RESULTS

After data cleaning, 33 242 bacterial isolates remained for analysis (Figure 1). The dataset included 25 848 and 7394 bacterial isolates cultured from dogs and cats respectively, with AST results representing 21 242 unique canine and feline patients during 2019–2022. Of 33 242 isolates during the 4-year period, 16 115 (49%) were gram-negative bacteria.

Of the 16 115 gram-negative isolates, 256 (1.6%) were resistant to imipenem, meropenem, or both. These 256 CROs represented specimens collected from 241 individual animals, specifically, 180 dogs and 61 cats. There were no differences in the overall prevalence between cats and dogs. Animals with CROs ranged in age from 0 to 17 years, with means of 7.3 years in dogs and 9.7 years in cats. Bacterial isolates cultured from early midlife animals (aged 5–9 years) represented 37.4% of CROs in dogs and 31.8% in cats (Table 1).

Table 1.

Gram-Negative and Carbapenem-Resistant Organisms From Animal Care Facilities in New York City, Identified by Culture and Antimicrobial Susceptibility Testing by a Single Commercial Diagnostic Laboratory, 2019–2022^a

Variable	Dogs		Cats		Total
	CROs, No. (%)b	GNOs, No.	CROs, No. (%)	GNOs, No.	CROs, No. (%)	GNOs, No.
All bacterial isolates	190 (1.6)	12 111	66 (1.6)	4004	256 (1.6)	16 115
Age of animal patientc
0–4 y (youth)	57 (1.9)	2961	6 (0.8)	732	63 (1.7)	3693
5–9 y (early midlife)	71 (1.8)	4024	21 (2.6)	802	92 (1.9)	4826
10–11 y (late midlife)	23 (1.2)	1976	15 (3.4)	445	38 (1.6)	2421
12–13 y (senior)	29 (1.6)	1767	12 (2.2)	550	41 (1.8)	2317
14–25 y (geriatric)	10 (0.8)	1332	12 (0.8)	1465	22 (0.8)	2797
Year of specimen submission
2019	38 (1.2)	3127	12 (1.2)	978	50 (1.2)	4105
2020	49 (1.9)	2548	17 (1.8)	961	66 (1.9)	3509
2021	52 (1.7)	3124	19 (2.0)	970	71 (1.7)	4094
2022	51 (1.5)	3312	18 (1.6)	1095	69 (1.6)	4407
Bacterial isolatesd
Bergeyella zoohelcum	0	2	1 (8.3)	12	1 (7.1)	14
Brevundimonas spp	1 (9.1)	11	0	3	1 (7.1)	14
Burkholderia spp	25 (55.6)	45	12 (44.4)	27	37 (51.4)	72
Chryseobacterium spp	6 (60.0)	10	5 (62.5)	8	11 (61.1)	18
Citrobacter spp	1 (1.6)	62	0	22	1 (1.2)	84
Elizabethkingia spp	6 (85.7)	7	2 (100.0)	2	8 (88.9)	9
Empedobacter spp	2 (100.0)	2	0	0	2 (100.0)	2
Enterobacter spp	20 (7.2)	278	8 (7.8)	102	28 (7.4)	380
Escherichia spp	28 (0.5)	5759	12 (0.5)	2513	40 (0.5)	8272
Klebsiella spp	54 (6.7)	804	15 (8.9)	169	69 (7.1)	973
Pantoea spp	1 (0.7)	136	0	22	1 (0.6)	158
Pseudomonas spp	42 (3.7)	1143	10 (3.6)	274	52 (3.7)	1417
Ralstonia spp	0	7	1 (50.0)	2	1 (11.1)	9
Sphingobacterium spp	3 (75.0)	4	0	0	3 (75.0)	4
Yersinia spp	1 (33.3)	3	0	0	1 (33.3)	3
Other	0	3838	0	848	0	4686
Specimen sitee
Nonsterile sites	175 (1.5)	11 726	56 (1.5)	3838	231 (1.5)	15 564
Abscesses	10 (3.1)	325	3 (2.1)	145	13 (2.8)	470
Ears	40 (2.6)	1527	4 (1.6)	243	44 (2.5)	1770
Lower respiratory tract	31 (10.8)	288	3 (2.6)	116	34 (8.4)	404
Upper respiratory tract	8 (4.9)	162	10 (3.3)	307	18 (3.8)	469
Urinary tract	35 (0.5)	7133	17 (0.7)	2601	52 (0.5)	9734
Skin and soft tissue	35 (2.1)	1657	9 (3.5)	257	44 (2.3)	1914
Surgical site infections	7 (7.9)	89	6 (15.8)	38	13 (10.2)	127
Other nonsterile sitesf	9 (1.7)	545	4 (3.1)	131	13 (1.9)	676
Normally sterile sitesg	8 (6.7)	120	1 (1.7)	58	9 (5.1)	178
Otherh	5 (4.1)	122	8 (14.8)	54	13 (7.4)	176
Unknown	2 (1.4)	143	1 (1.9)	54	3 (1.5)	197

Abbreviations: CROs, carbapenem-resistant organisms; GNOs, gram-negative organisms.

^aAll values were calculated after data deduplication, as described in Figure 1.

^bPercentages reflect the percentage positivity of CROs among all gram-negative isolates.

^cAge data were missing for 116 isolates.

^dA total of 54 gram-negative bacterial genera were isolated from cultures from dogs and cats during 2019–2022. Only gram-negative bacterial genera with ≥1 carbapenem-resistant isolate detected are listed.

^eA composite site variable was used for these data.

^fOther nonsterile sites included abscesses, gastrointestinal, ocular, oral, and reproductive tract sites.

^gNormally sterile sites included blood, cerebrospinal fluid, and samples taken from peritoneal, pericardial, and pleural sites, other internal body sites, joints, and bone.

^h“Other” included sites that could not be clearly categorized into the listed sites.

Detections ranged from 50–71 CRO isolates per year, with the percentage positivity ranging from 1.2% to 1.9% (Figure 2). The most frequently cultured CROs were Klebsiella spp, Pseudomonas aeruginosa, and Escherichia coli (Table 1). Some bacterial species showed carbapenem resistance in >50% of isolates, including Empedobacter spp (2 of 2 [100%]), Elizabethkingia spp (8 of 9 [89%]), Sphingobacterium spp (3 of 4 [75%]), Chryseobacterium spp (11 of 18 [61%]), and Burkholderia spp (37 of 72 [51%]).

Figure 2.

Number of isolates of carbapenem-resistant organisms (CROs) from animal care facilities in New York City, identified by culture and antimicrobial susceptibility testing by a single commercial diagnostic laboratory, 2019–2022, and percentage of gram-negative (GN) isolates testing positive as a CRO. Bars indicate total numbers of CRO isolates by companion animal species (dog or cat) and year. Dotted line shows the percentage positivity for carbapenem-resistance by year.

Most cultures from companion animals were sampled from nonsterile specimen sites (15 564 of 16 115 [96.6%]). However, of the 256 CROs, a higher proportion was detected in normally sterile sites (9 of 178 [5.1%]) than in nonsterile sites (231 of 15 564 [1.5%]) (Table 1). The most common specimen sites where CROs were detected were the urinary tract (52 of 256 [20.3%]), skin and soft tissue (44 of 256 [17.1%]), ears (44 of 256 [17.2%]), and the lower respiratory tract (34 of 256 [13.3%]) (Table 1). Carbapenem-resistant P aeruginosa and Burkholderia spp were most often cultured from ears, and carbapenem-resistant Klebsiella spp and E coli were most often cultured from urinary tract sources (Supplementary Table 3). Most of the CROs cultured from sterile sites (8 of 9 [88.9%]) were cultured from dogs. Only E coli, Enterobacter spp, and Klebsiella spp were cultured from normally sterile sites (Supplementary Table 3), and these were cultured from peritoneal fluid (5 of 9 [55.6%]), internal body sites (2 of 9 [22.2%]), and blood (1 of 9 [11.1%]). The remaining isolates were cultured from other sites (5.1%) or unknown (1.2%) sites.

CROs were not uniformly detected across NYC. Bacterial cultures were submitted from 85 of 313 NYC zip codes, with submissions from all 5 of the city's counties, also known as boroughs. Among these boroughs, New York County, also known as Manhattan, had the most gram-negative isolates cultured and the highest number of CRO isolates detected, more than three-quarters of all CROs detected from companion animals (Table 2). Zip code–level spatial heterogeneity was also observed: most zip codes had no CROs detected, 7 zip codes had 5–20 CROs, and 1 zip code had >100 CROs. Some organisms showed spatial patterning, while others did not. Despite being a common type of CRO, carbapenem-resistant P aeruginosa isolates did not show any spatial patterns, whereas 90% (62 of 69) of the carbapenem-resistant Klebsiella spp isolates and 86% (24 of 28) of carbapenem-resistant Enterobacter spp isolates were detected from a single zip code.

Table 2.

Carbapenem-Resistant Organisms From Animal Care Facilities in New York City, by Borough, Identified by Culture and Antimicrobial Susceptibility Testing by a Single Commercial Diagnostic Laboratory, 2019–2022^a

NYC Borough	NYC Zip Codes, No.		CROs Detected, No. (% of Total)b	GN Isolates Cultured, No.	Positivity of Isolates, %c
	With a Submitting Veterinarian	With a CRO Detected
Bronx	6	3	3 (1.2)	410	0.7
Brooklyn	22	11	24 (9.4)	3275	0.7
Manhattan	26	18	200 (78.1)	9473	2.1
Queens	25	11	23 (9.0)	1824	1.3
Staten Island	6	3	6 (2.3)	1133	0.5
All boroughs	85	46	256 (100)	16 115	1.6

Abbreviations: CRO, carbapenem-resistant organism; GN, gram-negative.

^aAll values are calculated after data deduplication, described in Figure 1.

^bPercentages reflect the number of CROs by borough relative to the total number of overall CROs detected (n = 256).

^cPercentage positivity reflects the number of CROs detected relative to total number of GN isolates submitted by veterinarians.

DISCUSSION

We aimed to establish a baseline understanding of CRO prevalence among companion animals in NYC by assessing gram-negative isolates from 1 commercial veterinary diagnostic laboratory with the goal to ultimately improve infection prevention and control. Collaborating with a commercial veterinary laboratory offered a high volume of specimens tested, a laboratory catchment area with submitting veterinarians in 85 NYC ZIP codes, and consistent AST laboratory methods.

Of the gram-negative isolates from companion animal specimens submitted by NYC veterinarians, 1.6% were CROs. The high number of CROs from Manhattan corresponded to a high number of specimens submitted and gram-negative isolates cultured, likely illustrating a preponderance of veterinary referral facilities and veterinarians submitting bacterial cultures within this borough. This distribution of animal CROs across the 5 boroughs differs from the human carbapenem-resistant Enterobacterales (CRE) surveillance data. Since 2018, human CRE laboratory reports are reportable to the health department in accordance with the NYC health code. While human CRE surveillance reports represent the patient’s borough of residence and do not include non-Enterobacterales organisms, such as Pseudomonas spp, during 2019–2022 there was a more uniform distribution of CRE cases across NYC, with 30% cases from Brooklyn, 27% from Queens, 17% from the Bronx, 15% from Staten Island, and 10% from Manhattan (New York City Department of Health and Mental Hygiene, 2025) [Unpublished data]. While the zip code of submitting veterinarian provides insight to potential facility-level transmission, the animal's borough of residence, which was unavailable in our dataset, would better enable comparing the geographic distribution of carbapenem-resistance among animals and humans to further assess other possible transmission routes (eg, household, community, etc).

The high number of Klebsiella spp and Enterobacter spp detected from a single NYC zip code likely represent ≥1 veterinary referral facilities in that area, since veterinary referral facilities are more likely to submit diagnostic bacterial cultures than general practice clinics. Referral facilities treat medically complex cases with animals in which first-line treatment has often failed, and animal patients from referral facility are more likely than the general animal population to be infected or colonized with CROs [23–26]. Without additional epidemiologic data or molecular characterization of the isolates, we cannot exclude the possibility of a local cluster or within-facility transmission. The increase in carbapenem-resistant P aeruginosa without a geographic focus could potentially be due to contaminated products [27]; however, similarly, this could not be determined without additional information. Above all, these results highlight that sharing data between public health and veterinary diagnostic laboratories can inform surveillance and ultimately might help limit the spread of CROs in animals and people.

Information on companion animal CROs has largely been gathered from outbreak investigations [23, 24, 27, 28]. To date, there have been 4 notable CRO outbreaks among companion animal in the United States, including clusters at 2 veterinary referral hospitals [23, 24], 1 at an animal rescue facility [28], and 1 associated with contaminated products [27]. CRO detection can be difficult as not all animals with a CRO become visibly sick. CRO colonization, when animals carry CROs in or on their bodies without showing signs of illness, can lead to environmental spread and direct transmission to other pets and people [29]. Spread via contaminated equipment in veterinary hospitals [24] and in community areas such as city parks [30] has been documented.

In addition to outbreak investigations, laboratory-based studies have also characterized CROs in animals. Veterinary Laboratory Investigation and Response Network (Vet-LIRN) laboratories in Kansas and Missouri had sporadic CRE companion animal detections over several years [3], and a nationwide study using commercial veterinary diagnostic data for Enterobacterales found that 0.76% of isolates were intermediate and 0.38% were resistant to imipenem [31].

Many animals with a CRO in our dataset were in the early midlife age category (aged 5–9 years), in contrast to CRO detections in humans, which most often occur later in life. A 2023 study of CRE in human patients across the United States reported a median age of 69 years, with most patients aged 65–79 years [32]. Age differences between human and animal populations could reflect differences between how humans and animals are exposed to CROs. For example, there is no animal equivalent to long-term care facilities, which are associated with CRO exposure in humans [7].

Common CROs in companion animals cultured from nonsterile and sterile sites are associated with different types of infections, with various causes. The most common CROs isolated from companion animals, Klebsiella spp, E coli, and P aeruginosa, were most often isolated from nonsterile sites, which could indicate infections by normal flora or contamination of a damaged body site, such as a skin abrasion [33]. These bacteria can be the cause of urinary tract infections, pyoderma, and otitis externa [34, 35], and the use of empiric antibiotic treatment for these conditions is common [36, 37]. In contrast, the most common organisms cultured from normally sterile sites included E coli, Enterobacter spp, and Klebsiella spp, and peritoneal fluid was the most common sterile site for CRO isolation. Peritonitis in companion animals may result from various causes, such as hematogenous spread or damage to the gastrointestinal or urogenital tract [38].

Some less commonly cultured genera showed a high proportion of carbapenem resistance, including Empedobacter spp, Elizabethkingia spp, Sphingobacterium spp, Chryseobacterium spp, and Burkholderia spp; Elizabethkingia spp, and Chryseobacterium spp have been described as emerging infections in human hospitals and long-term care settings [39, 40], whereas Burkholderia spp are water-related opportunistic organisms historically known to be associated with environmental contamination within healthcare settings and nosocomial infections [41, 42]. In companion animal veterinary settings, Burkholderia spp have been found in contaminated chlorhexidine 2% scrub and associated with nosocomial infections in cats [43]. Given the sporadic detections of these genera, high proportion of carbapenem resistance observed, and their potential for nosocomial spread, veterinarians and public health partners should carefully review AST data and consider the potential for within-hospital transmission when these organisms are cultured.

Our findings are subject to at least 5 limitations. First, we assessed CROs among gram-negative isolates cultured from specimens submitted by veterinarians. This likely overestimated CROs in the general population of companion animals residing in NYC, as animals with recurrent or chronic conditions are more likely to have AST performed [36, 37]. Second, our study period overlapped with the emergency phase of the coronavirus disease 2019 (COVID-19) pandemic. Although there were anecdotal reports of increases in demand for veterinary services during the COVID-19 pandemic [44], there is limited available literature to know whether this increase was widespread or sustained or whether it fluctuated and how this may have affected CRO detections during 2020–2022.

Third, high veterinary costs may further limit testing since bacterial cultures are typically paid for by pet owners, potentially overrepresenting pets living in households with a higher socioeconomic status. Lower socioeconomic communities have disproportionately been affected by resistant bacteria, yet pets from households unable to pay for bacterial cultures were not included in our dataset [45]. Fourth, our data do not include complete AST or whole-genome sequencing data, and the source was a single commercial veterinary diagnostic laboratory and thus does not represent sampling at all animal care facilities in NYC. While we cannot quantify the number of facilities that use this laboratory, it is considered 1 of 2 veterinary laboratories that dominate the industry [46]. Fifth, because the geographic resolution of our dataset was the zip code of the submitting animal care facility, we were unable to distinguish whether prevalence increases within a zip code were attributable to >1 facility.

In animal populations, delayed detection of emerging and zoonotic pathogens and limited notification to public health entities can ultimately result in increased transmission between animals or from animals to humans [47]. While this report and other companion animal CRO studies are retrospective [3, 31], for public health entities to prioritize and implement interventions to control CROs in veterinary settings, systematic animal CRO surveillance systems are ideally needed to provide real-time CRO notifications, resistance mechanism testing, and facility-level data. Strengthening CRO surveillance in animals, especially in collaboration with veterinary diagnostic laboratories, could improve understanding and enable public health action to help control the spread of CROs in companion animals.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.