Patterns of antimicrobial, multidrug and methicillin resistance among Staphylococcus spp. isolated from canine specimens submitted to a diagnostic laboratory in Tennessee, USA: a descriptive study

Background Multidrug- and methicillin-resistant staphylococci are both veterinary and public health concerns due to their zoonotic potential. Therefore, the objective of this study was to investigate patterns of antimicrobial, multidrug, and methicillin resistance among four Staphylococcus spp. commonly isolated from canine clinical specimens submitted to the Clinical Bacteriology Laboratory at the University of Tennessee College of Veterinary Medicine (UTCVM). Methods Results of antimicrobial susceptibility testing and mecA polymerase chain reaction (PCR) for isolates of four common Staphylococcus spp. isolates were obtained from the Bacteriology Laboratory at the UTCVM between 01/01/2006 and 12/31/2017. Cochran-Armitage trend test was used to assess temporal trends of antimicrobial resistance (AMR), multidrug resistance (MDR), and methicillin resistance. Kappa test of agreement was used to assess agreement between the results of PCR and disk diffusion tests. Results Most of the 7805 isolates were S. pseudintermedius (6453 isolates), followed by S. coagulans (860), S. aureus (330), and S. schleiferi (162). Among S. pseudintermedius isolates, 45.5% were MDR, and 30.8% were methicillin-resistant (MRSP). There was a significant temporal increase in MRSP (p = 0.017). Chloramphenicol resistance increased among both MRSP and methicillin-susceptible (MSSP) isolates (p <  0.0001). Among S. aureus isolates, 40.9% were MDR, 37.4% were methicillin-resistant (MRSA), and the proportion of MRSA isolates increased significantly (p = 0.0480) over time. There was an increasing temporal trend in the proportion of MDR isolates among MSSP (p = 0.0022), but a decrease among MRSP (p <  0.0001) and MRSA (p = 0.0298). S. schleiferi had the highest percentage (56.9%) of methicillin-resistant isolates. Oxacillin disk diffusion was superior to cefoxitin for the detection of mecA-mediated resistance and had almost perfect agreement with mecA PCR assay for S. pseudintermedius (95.4% agreement, kappa (κ) = 0.904; p <  0.0001), S. coagulans (95.6%, κ = 0.913; p <  0.0001) and S. schleiferi (97.7%, κ = 0.945; p <  0.0001). However, cefoxitin disk diffusion was superior to oxacillin disk diffusion and had almost perfect agreement with mecA PCR assay for S. aureus (95.3%, κ = 0.834; p <  0.0001). Conclusions The levels of resistance and increasing temporal trends are concerning. These findings have implications for treatment decisions and public health due to the zoonotic potential of staphylococci. Continued surveillance and use of antibiograms to guide clinical decisions will be critical.

Numerous studies have reported that most canine Staphylococcus spp. isolates display resistance to at least one antimicrobial [6,14,20,21]. Varying levels of multidrug resistance (MDR), defined as resistance to at least one drug in three or more classes of antimicrobials [22,23], have also been detected among staphylococci isolated from companion animal specimens [14,20,21,24,25]. Moreover, methicillin resistance, which has received major attention in human medicine as the spread of hospital-and community-associated methicillin-resistant S. aureus (MRSA) challenges infection control efforts worldwide, is also increasingly being detected among canine Staphylococcus isolates, particularly S. pseudintermedius [26][27][28][29][30][31].
The occurrence of MDR and methicillin resistance among canine Staphylococcus isolates has important clinical implications. Importantly, infections with resistant isolates may require treatment with antimicrobials considered high priority for human health [32,33]. Methicillin resistant staphylococci (MRS) express the mecA gene or its homolog, mecC, and produce an altered penicillin binding protein (PBP), which confers resistance to all β-lactam antibiotics, severely limiting treatment options [31,[34][35][36]. Therapeutic failures resulting from infections with resistant isolates also present serious clinical challenges to clinicians and increase the financial burdens for animal owners [37,38]. Furthermore, evidence suggesting cross-transmission of resistant pathogens between animals and humans implies that the expansion of these organisms is also a public health concern [39][40][41][42]. S. pseudintermedius has been identified in human clinical infections, and previous reports suggest that pet owners are at risk of zoonotic transmission [43,44]. In addition, while S. aureus is mainly a pathogen of humans, pets exposed to methicillin resistant S. aureus (MRSA) infections by their owners may play a role in perpetuating MRSA colonization in the household [45]. The potential for exposure to antimicrobial or multidrug resistant Staphylococcus isolates among those that are immune-compromised is of particular concern [46].
The existence of AMR, MDR, and methicillin resistance among canine Staphylococcus isolates, coupled with the public health implications of zoonotic transmission of these organisms, highlights the importance of monitoring antimicrobial susceptibility patterns. Temporal changes in antimicrobial resistance patterns among canine staphylococcal isolates imply that continued surveillance is essential to identify patterns, trends, and newly emerging resistance [14,47]. Geographically relevant and timely epidemiologic monitoring is useful for the development of evidence-based antimicrobial use guidelines to inform veterinarians of current best practices. It is also important that surveillance data regarding regional AMR patterns for patients from a broad range of clinical settings, rather than specialty practices alone, are available to practitioners to enable informed antimicrobial selection when empirical therapy is necessary. Therefore, the objective of this study was to investigate patterns of antimicrobial, multidrug, and methicillin resistance among four clinically important Staphylococcus species (S. pseudintermedius, S. aureus, S. coagulans, and S. schleiferi) isolated from canine specimens submitted to the Clinical Bacteriology and Mycology Laboratory at the University of Tennessee College of Veterinary Medicine by both referring veterinarians and teaching hospital personnel between 2006 and 2017.

Study design and data source
This descriptive retrospective study used laboratory records of canine clinical specimens processed at the Clinical Bacteriology Laboratory at the University of Tennessee College of Veterinary Medicine (UTCVM) between January 1, 2006 and December 31, 2017. Records for a total of 7810 canine clinical specimens positive for S. pseudintermedius, S. aureus, S. coagulans, or S. schleiferi were available for analysis. The data were assessed for duplicate entries and five were identified and removed, leaving 7805 unique isolates for subsequent analyses. The following data were extracted from the laboratory records: patient identification, medical record number, specimen collection site, Staphylococcus species isolated, results of antimicrobial susceptibility tests and mecA polymerase chain reaction (PCR) assay. Hospital patient records for the study period were also obtained from the UTCVM veterinary teaching hospital and matched to the laboratory records. The following data were extracted from the hospital records: medical record number, patient type (hospital patient specimen vs. submission by referring veterinarian), age, sex, species, and breed.

Bacterial isolation, antimicrobial susceptibility testing, and PCR assay
Bacterial isolation and antimicrobial susceptibility tests were performed according to standard practice using previously described methods [47]. Briefly, Staphylococcus spp. isolates were identified with a biochemical identification process, using the following tests: tube coagulase, phenol red broth with lactose and with trehalose, and the Voges-Proskauer test [48]. The diagnostic laboratory followed Clinical and Laboratory Standards Institute (CLSI) standards for antibiotic susceptibility testing [49][50][51][52][53]. Antimicrobial susceptibility tests for the majority of bacterial isolates were performed using the disk diffusion method. However, some were tested using a microbroth dilution method using an automated susceptibility testing system [47,54]. Antimicrobial agents from the following classes were included in the standard antimicrobial susceptibility panel for Staphylococcus spp.: β-lactams (ampicillin, amoxicillin/clavulanic acid, cefoxitin, cefpodoxime, cephalothin, oxacillin and penicillin), fluoroquinolones (marbofloxacin), folate inhibitors (trimethoprim/sulfamethoxazole [TMS]), lincosamides (clindamycin), macrolides (erythromycin), phenicols (chloramphenicol), and tetracyclines (tetracycline). Reporting of susceptibility to β-lactam antibiotics in the present study was limited to ampicillin, oxacillin, and cefoxitin. Ampicillin (rather than penicillin) was used to represent susceptibility to penicillinase-labile penicillins in this study because the results of ampicillin testing were available for a larger number of Staphylococcus spp. isolates. For isolates deemed to be resistant to multiple drug classes, susceptibility tests for the following agents were performed in addition to the standard panel: amikacin, doxycycline, minocycline, and rifampin. Amikacin susceptibility tests were also performed for all cutaneous isolates, while gentamicin was used for otic and urinary isolates.
Oxacillin zone sizes of ≤17 mm (resistant) and ≥ 18 mm (susceptible) were used as breakpoints for classification of S. pseudintermedius, S. coagulans, and S. schleiferi [55,56]. In addition, interpretive criteria recommended for coagulase-negative Staphylococcus spp. for cefoxitin were followed for S. pseudintermedius, S. coagulans, and S. schleiferi [49][50][51][52][53]. Otherwise, the laboratory followed CLSI standards in place during the year of specimen submission, and isolates were classified as susceptible, intermediate, or resistant based on these criteria [49][50][51][52][53]. During 2006, a real-time PCR assay was used for the detection of the mecA gene for all Staphylococcus spp. isolates, with a cycle threshold value of ≤30 being considered positive [47,57]. During all remaining years of the study period, conventional PCR was employed for this purpose, using previously described methods [55]. After 2006, mecA PCR was performed for all staphylococcal isolates resistant to oxacillin or cefoxitin, as well as for 100 oxacillin-susceptible S. pseudintermedius isolates annually.

Data management
Data management and statistical analysis were performed using SAS Version 9.4 [58]. Medical record numbers were used to join patient data extracted from hospital records to the laboratory records. Patient age was categorized into the following groups: < 3 years of age, 3-6 years, 6-9 years, 9-12 years, and ≥ 12 years. Patient sex categories included: intact male, castrated male, intact female, and spayed female. Patient breed was placed into one of the following categories based on American Kennel Club (AKC) designations: herding, hound, toy, non-sporting, sporting, terrier, and working breed groups [59]. Patients with multiple breeds were coded as "mixed breed, " and those with a listed breed not recognized by the AKC were coded as "other". Specimen collection sites were categorized as ear, skin, urine or bladder, joint or bone, mucosa, and "other".
Results of antimicrobial susceptibility tests were excluded from analysis if differing interpretations were obtained upon repeated testing. Thus, results for the following agents were excluded from analysis: amikacin (2 isolates), oxacillin (6 isolates), cefoxitin (2 isolates), and cefpodoxime (2 isolates). Results of mecA PCR assay were excluded from analysis for 14 isolates for which repeated test results had differing interpretations. Therefore, results of mecA PCR assay were included for a total of 4152 isolates. All antimicrobial susceptibility test results were re-classified into two groups, susceptible and resistant, with the latter encompassing all non-susceptible isolates (those classified as "non-susceptible, " "intermediate, " or "resistant") [23].
The antimicrobial agents used for susceptibility testing were categorized into the appropriate drug classes. Isolates that were non-susceptible to at least one agent in one or more antimicrobial classes, excluding intrinsic resistance, were classified as AMR [22,23]. Isolates that were non-susceptible to one or more agents in 3 or more antimicrobial classes, excluding intrinsic resistance, were classified as MDR [22,23]. Isolates were classified as methicillin-resistant based upon results of phenotypic susceptibility testing, confirmed with positive PCR results for the mecA gene.

Statistical analysis
The Kolmogorov-Smirnov test was used to assess for normality of distribution of patient age, which was found to be non-normally distributed and hence median and interquartile range (IQR) were computed. The following categorical variables were assessed: age group, patient breed category, patient sex, patient type, and specimen collection site. The Cochran-Armitage trend test was used to assess for significant temporal trends in AMR, MDR, and methicillin resistance, using a two-sided p-value with a cutoff of ≤0.05. This analysis was performed for S. pseudintermedius, S. aureus, S. coagulans, and S. schleiferi isolates. In addition, separate analyses were conducted for methicillin-resistant and methicillin-susceptible S. pseudintermedius (MRSP and MSSP) and S. aureus (MRSA and MSSA) isolates. To facilitate comparison between methicillin-resistant and methicillin-susceptible isolates, temporal trends in antimicrobial and multidrug resistance were displayed in graphs, which were generated using R version 4.1.1 [60]. Chi-square tests (or Fisher's exact tests where appropriate due to sample size) were used to assess for significant differences in levels of antimicrobial and multidrug resistance between methicillinresistant and methicillin-susceptible isolates. Kappa tests of agreement were used to assess agreement between the results of mecA PCR assay and oxacillin as well as cefoxitin susceptibility tests.

Summary statistics
A total of 7805 canine clinical specimens with single isolates that met the inclusion criteria for the study were assessed. The most common species was S. pseudintermedius, with 6453 isolates, followed by S. coagulans (860 isolates), S. aureus (330 isolates), and S. schleiferi (162 isolates).

b) S. aureus
Resistance to ampicillin was observed among the majority (82.6%) of S. aureus isolates (Table 3). Low levels of resistance were observed for TMS (4.2%), chloramphenicol (2.4%), gentamicin (1.1%), and tetracycline (6.4%), while resistance to clindamycin and erythromycin were relatively high (35.4 and 47.3%, respectively). The overall percentage of MRSA isolates was 37.4%, and increased significantly during the study period (p = 0.0480), from 21.6 to 40%. In addition, 40.9% of S. aureus isolates were MDR. Among the subsets of isolates tested for susceptibility to amikacin and rifampin, several exhibited resistance to amikacin (12/78), but none were resistant to rifampin (0/21). No significant temporal trends were observed for these agents.

iii) S. coagulans
Ampicillin (40.9%) resistance among S. coagulans isolates was comparatively lower than that of S. pseudintermedius and S. aureus (Table 4). Resistance to gentamicin (36.2%) and marbofloxacin (37.7%) were relatively common among S. coagulans isolates, while resistance to TMS, clindamycin, erythromycin, and tetracycline were rarely observed, and none exhibited resistance to chloramphenicol. Among the subsets of isolates tested for susceptibility to amikacin (232 isolates) and rifampin (28 isolates), none exhibited resistance.
The percentages of AMR (57%) and MDR (20%) S. coagulans isolates were much lower than those for S. pseudintermedius and S. aureus. Furthermore, MDR showed a significant (p = 0.0004) decreasing temporal  Methicillin-resistant S. pseudintermedius, 2 Methicillin-susceptible S. pseudintermedius, 3 Trimethoprim/sulfamethoxazole, 4 Multidrug resistance. *Indicates that upper limit of y-axis has been reduced to improve visual comparison trend, with the percentage of resistant isolates changing from 21.7% at the beginning of the study period to 11.8% at the end. On the other hand, overall methicillin resistance was similar to S. pseudintermedius and S. aureus (32.2%), but did not exhibit any significant temporal trends.

iv) S. schleiferi
Similar to findings for S. coagulans, S. schleiferi isolates also exhibited a lower level of resistance to ampicillin (56.5%) than S. pseudintermedius and S. aureus (Table 5). However, S. schleiferi isolates had the highest level of methicillin resistance of all species in this study (56.9%). Approximately half of the S. schleiferi isolates showed resistance to gentamicin (45.5%) and marbofloxacin (48%), while resistance to other drugs, including TMS (2.5%), clindamycin (8.6%), erythromycin (8.6%), and tetracycline (6.8%) was much less common. However, resistance to tetracycline did increase significantly during the course of the study period (p = 0.0217). Among the subsets of isolates tested for susceptibility to amikacin (56 isolates) and rifampin (7 isolates), resistance to amikacin was observed in a single isolate, and none were resistant to rifampin.
The majority of S. schleiferi isolates were resistant to at least one antimicrobial (77.2%), but MDR was again comparatively lower than the other Staphylococcus species (27.2%). Methicillin resistance fluctuated but did not exhibit any significant temporal trends.

Methicillin resistance and mecA gene detection a) S. pseudintermedius
A total of 97.4% of the cefoxitin-resistant S. pseudintermedius isolates (based on disk diffusion test) were positive for the mecA gene (based on PCR assay), while only 62.1% of cefoxitin-susceptible isolates were negative for mecA (Table 6). Overall, there was 74.9% agreement between cefoxitin disk diffusion testing and mecA PCR assay. The kappa test of agreement indicated moderate agreement (kappa = 0.522; p < 0.0001) between the two methods. In contrast, oxacillin disk diffusion test had almost perfect agreement with mecA PCR (95.4% agreement; κ = 0.904; p < 0.0001).

Discussion
This study assessed patterns of antimicrobial, multidrug and methicillin resistance among Staphylococcus isolates from canine clinical specimens submitted to the UTCVM Bacteriology Laboratory between 2006 and 2017. The majority of these staphylococci were isolated from skin, ear and mucosal specimens, reflecting their importance as causative agents of pyoderma and otitis [3][4][5][6], which are frequently encountered in clinical practice. Surveillance of antimicrobial susceptibility patterns among canine Staphylococcus isolates is therefore valuable for guiding evidence-based treatment decisions, and useful for identifying trends that may reflect selection pressure. The majority of S. pseudintermedius isolates were resistant to ampicillin (84.3%) and therefore most were classified as AMR, consistent with previous reports [28,47,61,62]. Indeed, resistance to penicillinaselabile penicillins has become widespread over the past several decades [63], limiting their usefulness for the treatment of staphylococcal infections in dogs.
Relatively high levels of MRSP were observed throughout the study period (30.8% overall), with a mild but statistically significant temporal increase. This follows a steep increase in mecA-mediated resistance among isolates submitted to the UTCVM bacteriology laboratory between 2001 and 2007 [55]. These findings are not unprecedented, as mecA-mediated resistance has emerged among S. pseudintermedius in North America and Europe [64,65], and MRSP has been detected around the world [66][67][68][69][70][71]. However, the level of methicillin resistance among isolates from canine specimens in the current study was higher than previously observed in other locations within the U.S., such as Kentucky [14] and Michigan [72]. These regional differences imply that geographically relevant AMR surveillance provides valuable information for veterinary practitioners.
In addition to the substantial level of methicillin resistance, almost half of the S. pseudintermedius isolates in the current study were MDR. Though a temporal decline in the proportion of MDR MRSP isolates was observed, the vast majority of MRSP (90.5%) were MDR. A limited number of antimicrobials are available to effectively treat MDR and MRSP infections; these isolates are often resistant to drug classes used in veterinary medicine such as tetracyclines, fluoroquinolones, macrolides, trimethoprim-sulfonamides, and lincosamides [73]. Considerable levels of resistance to the above agents (exceeding 75%) were observed among MRSP isolates in the current study.
As a result of such limitations in available treatment options, infections with MDR and/or MRSP isolates may necessitate the use of antimicrobials such as rifampin or chloramphenicol, despite the potential for adverse effects [73][74][75][76]. Resistance to rifampin was only assessed among a subset of MDR isolates in the current study, and was infrequently observed. On the other hand, it is concerning that a statistically significant increase in resistance to chloramphenicol occurred among both MRSP and MSSP isolates during the study period, with the highest levels among MRSP (from 4.0 to 12.0%). This finding is particularly concerning given its utility for the treatment of MRSP infections. A similar but accelerated trend was reported among canine isolates collected at the Texas A&M Veterinary Teaching Hospital, likely due to selection pressure from the increased use of chloramphenicol in the face of MDR/MRSP infections [77]. However, while that study also reported an increase in resistance to amikacin [77], another potential treatment for MDR or MRSP infections [73], resistance to amikacin was rare among MDR and cutaneous isolates in the current study.
In addition to the clinical challenge associated with the treatment of MDR and MRSP infections in veterinary patients, the zoonotic potential of S. pseudintermedius raises concerns from a public health perspective. Human infections with S. pseudintermedius are typically associated with companion animal contact [43,[78][79][80][81][82], and have included MRSP infections [83,84]. Given the potential for rapid selection of resistant organisms, ongoing surveillance is warranted to ensure prompt recognition of clinically relevant patterns of AMR and MDR. Indeed, while higher levels of resistance to non-β-lactam antibiotics among MRSP compared to MSSP are expected [63], the temporal increase in MDR MSSP isolates observed in the current study was an interesting, albeit concerning finding. This trend appears to have been driven by rising levels of resistance to fluoroquinolones, folate inhibitors, lincosamides, and macrolides, all of which exhibited decreases for MRSP isolates.

b) S. aureus
S. aureus was less commonly isolated in comparison to S. pseudintermedius in this study, as dogs are not the primary reservoir for this organism [85]. In the current study, 37.4% of S. aureus isolates were methicillin-resistant, and MRSA increased substantially during the study period (from 21.6 to 40%). Varying levels of methicillin resistance have been reported among canine clinical S. aureus isolates in other recent studies, ranging from 12.8% in an Australian study [86], to 62.7% of isolates from wound infections in a German study [87]. S. aureus strains in companion animals tend to mirror those circulating in humans [87,88], and canine colonization and infections with S. aureus often result from reverse zoonosis [45,[89][90][91]. However, pets may serve as reservoirs once exposed, and MRSA carriage in humans may be difficult to eliminate if household dogs harboring the organism are not treated [45,90]. The susceptibility patterns of S. aureus differed somewhat from S. pseudintermedius, which may reflect antimicrobial use patterns, particularly if humans are a major source of S. aureus colonization in this population of dogs. For example, resistance to gentamicin, tetracycline, and TMS were much less common among S. aureus, and a temporal trend was not observed for chloramphenicol. Furthermore, the proportion of MDR isolates did not increase among MSSA. However, a more rapid increase in methicillin resistance was observed among S. aureus isolates compared to S. pseudintermedius. Further investigation is warranted to understand the epidemiology of S. aureus in this patient population and identify factors which may have contributed to the observed patterns, including the pronounced increase in MRSA.

iii) S. coagulans and S. schleiferi
While S. coagulans and S. schleiferi (formerly S. schleiferi subsp. coagulans and S. schleiferi subsp. schleiferi [15]) have received increased attention for their roles as human and animal pathogens in recent years, there are fewer published reports of AMR patterns for these organisms in comparison to S. pseudintermedius and S. aureus. The percentage of methicillin resistant isolates was the most notable difference between S. coagulans and S. schleiferi (32.2 and 56.9%, respectively), a finding which has previously been reported [19]. In fact, S. schleiferi isolates had the highest proportion of methicillin resistant isolates in the current study, consistent with previous reports [19,92]. In addition to being isolated from healthy dogs and first-time infections, S. coagulans and S. schleiferi may be particularly important in recurrent infections such as otitis and pyoderma [17,18,[93][94][95]. Prior exposure to β-lactam antibiotics, commonly used for empirical treatment of skin infections, is a reported risk factor for oxacillin (methicillin) resistance among S. coagulans and S. schleiferi isolates from canine specimens [19].
Despite the considerable proportion of methicillinresistant S. schleiferi isolates, several other antimicrobial classes appear to remain as potential treatment options.  [52]. For isolates that did not have borderline zone sizes and/or MIC values, the possibility of an alternate mechanism of oxacillin resistance cannot be excluded based on the information available.
Nonetheless, results of the current study support the usefulness of oxacillin disk diffusion for detecting mecAmediated resistance in S. pseudintermedius, S. coagulans, and S. schleiferi.
For S. aureus, cefoxitin was superior to oxacillin disk diffusion for detecting mecA-mediated resistance, consistent with CLSI recommendations [52,98]. While there was almost perfect agreement between the two tests, discordant results were obtained for seven S. aureus isolates. In order to facilitate comparison between studies, CLSI-recommended interpretive breakpoints in place at the time of sample submission were used in the current study. The recommended cutoff values for cefoxitin changed during the study period, from ≤19 mm (resistant), ≥20 mm (susceptible) to ≤21 mm (resistant), ≥22 mm (susceptible) [49,50]. Under current CLSI standards, one of the three mecA-positive, cefoxitin-susceptible isolates would have been re-classified as cefoxitin-resistant, resulting in agreement between the two tests.
Further investigation to determine the mechanism of phenotypic methicillin resistance for the mecA-negative Staphylococcus spp. isolates was beyond the scope of the present study. However, the identification of this pattern among several Staphylococcus spp. isolates in this study suggests that PCR assay for both the mecA and mecC genes may be warranted in future research.

Study strengths and limitations
The identification of antimicrobial resistance patterns and trends using a large sample of canine patients in this study are useful for informing evidence-based treatment decisions. In addition, clinical specimens were obtained from submissions by referring veterinarians in addition to hospital patients, representing a broad range of patients from both primary practice and specialty services. Antimicrobial susceptibility data were reported at the level of individual drugs, allowing for a detailed assessment of these patterns.
However, this study is not without limitations. For instance, while Staphylococcus intermedius group (SIG) species cannot be reliably distinguished by biochemical identification processes, virtually all canine SIG isolates are S. pseudintermedius; therefore, as recommended, SIG isolates were reported as S. pseudintermedius [100, 101]. S. coagulans and S. schleiferi were differentiated based upon tube coagulase testing, which may be inconsistent with genotypic testing [102], and therefore some of these isolates may have been misclassified. In addition, patient medical history, including current or prior antimicrobial therapy, was not available, precluding our ability to identify drivers of the observed patterns of AMR and MDR.
Finally, the use of specimens submitted for culture may present some selection bias, as these are unlikely to be from first-time infections, and may have been submitted due to therapeutic failure. However, the majority of specimens were from submissions by referring veterinarians, and samples were not solely obtained from patients of specialty services, resulting in a broader picture of AMR patterns across a variety of clinical settings. Despite the above limitations, the findings of the current study provide valuable insights to patterns and trends of AMR in the region, and are useful for guiding both clinical decisions and directions for future research.

Conclusion
There is evidence of substantial levels of MDR among S. pseudintermedius and S. aureus isolates from canine specimens, and increasing temporal trends in MRSP and MRSA. Unfortunately, this suggests that few agents remain broadly effective against these organisms. Furthermore, the emergence of chloramphenicol resistance observed in the current study is concerning. Future studies will investigate potential drivers of the observed antimicrobial susceptibility patterns. The observed levels of MDR and methicillin-resistance highlight the importance of antibiograms to guide treatment decisions. The temporal increase in canine methicillin-resistant staphylococci observed in this study has practical implications for human health. These findings highlight the need for continued surveillance of antimicrobial resistance patterns among staphylococci, and the importance of judicious antimicrobial use practices.