Results of urinary bacterial cultures and antibiotic susceptibility testing of dogs and cats in the UK

Objectives Bacterial urinary tract infections are a common diagnosis in small animal practice and antibiotics are often administered empirically. The aim of this study was to investigate the aetiology and antibiotic resistance of uropathogens in dogs and cats in the UK. Materials and Methods Retrospective study of uroculture and antibiotic susceptibility testing results (n=808) by disk diffusion processed at a veterinary pathology laboratory between 2011 and 2012. Results Significant bacteriuria was detected in 18.4% of samples from dogs and 10.0% from cats, most of which (>90%) yielded a single organism. Escherichia coli was the most prevalent bacterial species (54.7% and 55.6% of feline and canine isolates, respectively) followed by Proteus mirabilis in dog samples (22.7%) and Enterococcus spp. in cat samples (23.2%). Approximately a third of E. coli isolates were resistant to ampicillin but resistance was much lower among Enterococcus spp. and P. mirabilis. Resistance to amoxicillin‐clavulanic acid also seemed to be emerging, particularly in E. coli (almost 20% resistant). In contrast, resistance to trimethoprim‐sulfamethoxazole for uropathogens remained <13% except for P. mirabilis (19.4%). Overall, fluoroquinolones showed the best in vitro activity (resistance mostly below 10% for enrofloxacin and marbofloxacin). Clinical Significance Our results provide evidence of the emergence of resistance to antibiotics commonly used to treat bacterial urinary tract infections. Continued monitoring of the patterns of antibiotic resistance in uropathogens is needed to assess the adequacy of recommendations on the empiric therapy of these infections.


INTRODUCTION
Bacterial urinary tract infections (UTIs) are diagnosed in approximately 14% of dogs during their lifetime (Ling 1984). UTIs are more common in cats aged 10 or above, those with concurrent diseases (e.g. diabetes mellitus, chronic renal failure) and those that have undergone urinary catheterisation or perineal urethrostomy (Lulich et al. 1992, Litster et al. 2007. UTIs can cause a spectrum of disease from asymptomatic to life threatening. The clinical signs known to be compatible with UTI include dysuria, stranguria, pollakiuria and hematuria, but such signs are not pathognomonic of infection. A diagnosis of UTI can be aided by the detection of significant bacteriuria in quantitative cultures of urine specimens (Weese et al. 2011, Olin & Bartges 2015. In the same manner, resolution of clinical signs is not a sign of infection elimination and cure (especially in chronic or recurrent UTIs) and can only be determined with urine culture, ideally 5 to 7 days after therapy is terminated (Barsanti & Finco 1979, Ball et al. 2008, Olin & Bartges 2015. Urine samples for microbiological examination should ideally be collected from all patients suspected of bacterial UTI before the initiation of antibiotic treatment to confirm the diagnosis and determine the aetiology of the infection. Empirical treatment, while laboratory results are pending, is usually recommended to relieve patient discomfort (Weese et al. 2011). An aetiological diagnosis may not be pursued due to financial constraints of the pet owner and/or a good initial response to empirical antibiotics. Consequently, continuation with the selected treatment occurs without any testing. Current knowledge on the local susceptibility patterns of the most common uropathogens is therefore necessary to ensure adequate empiric antibiotic coverage. Antibiotic choices should also take into account the therapeutic levels reached in urine; therefore antibiotics with high renal excretion rates are preferred (Kramer et al. 2012). Inadequate empirical choices can lead to poor patient outcomes and contribute to the selection of resistance.
The emergence of multidrug-resistant (MDR) UTIs, which has already been reported (Cohn et al. 2003, Penna et al. 2010, can seriously complicate the treatment of these infections by limiting the effective drugs available. There are also public health concerns associated with the potential zoonotic transmission of some of these resistant organisms, which is facilitated by the close proximity between humans and their household pets (Starcic et al. 2002, Guardabassi et al. 2004, Weese 2008, Rampacci et al. 2018.
The objectives of the present study were to provide evidence on the aetiology and antibiotic resistance of bacterial uropathogens in dogs and cats in the UK using retrospective data from referrals to a diagnostic laboratory.

Data collection
Data from urine samples from dogs and cats with significant bacteraemia submitted to NationWide Laboratories (NWL; Lancashire, UK) for microbiological examination between November 2011 and November 2012 were retrieved from electronic records. The geographic distribution of the veterinary practices from which the samples were originated is shown in Fig 1. Information on the method of urine collection was not provided by the requesting veterinarian. Urine samples of an animal were only submitted once (non-repetitive samples). The collected data included age and gender of the patients and the results from uroculture and antibiotic susceptibility testing. A cut-off of ≥10 5 colony-forming units (CFU)/mL and growth of  (Sørensen et al. 2016). Only samples that matched these criteria were included in the analysis and if more than two isolates were cultured this was considered contaminated. Information on the clinical status of the patient (signs of a UTI or subclinical infections) was not provided but samples were submitted for clinical purposes, e.g. clinical suspicion of a UTI or screening susceptible patients.

Laboratory methods
Semi-quantitative bacterial cultures of undiluted urine samples were performed by the calibrated loop method. A 10-μL loop was used to inoculate the urine onto 5% blood agar and Mac-Conkey agar. The plates were placed at 37°C and checked for growth after 18 to 24 hours of incubation. Bacteria were identified by Gram stain and standard biochemical procedures. Antimicrobial susceptibility testing was done by the Kirby-Bauer method according to the recommendations of the Clinical and Laboratory Standards Institute (CLSI 2008). Quality control strains (Staphylococcus aureus ATCC 25923, Escherichia coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853) were prepared and tested by exactly the same methods as employed for the test isolates. NWL where the microbiological testing was carried out held an ISO17025 accreditation for clinical microbiology testing of urine. The antibiotics tested were penicillin, ampicillin, amoxicillin-clavulanic acid, cephalothin, cefovecin, tetracycline, enrofloxacin, marbofloxacin, erythromycin and trimethoprim-sulfamethoxazole. Clinical breakpoints defined by CLSI (CLSI 2008) were applied and noted differences in breakpoint data between CLSI (2008) and CLSI VET01S (2020) are included in Table 1.

Data analysis
Data records were exported in a single batch from the Laboratory Information system by NWL (Winpath, Clinisys) using the application's query function. All records with a species code "canine" or "feline" and a test request code "urine culture" were included in a comma-separated format that was then imported into Microsoft Excel. Results were further analysed by a single operator. Results were expressed as percentages. The frequency of different bacterial species in urine specimens was calculated by dividing the number of isolates from a certain species by the total number of isolates of this species from the canine and feline samples. For each antibiotic, the percentage of resistance was calculated by dividing the number of resistant isolates by the total number of isolates reported.  Table 2. More than half of the samples from dogs and cats (808) belonged to female dogs (52.7%; n=426); for cats this figure amounted to 10.4% (n=84).

Bacterial isolates
Most urine specimens from both cats (94.2%) and dogs (94.3%) yielded the pure growth of a single organism. Two different isolates were present in the remaining samples (approximately 6% for cat and dog samples each). Gram-negative species predominated, particularly those belonging to the Enterobacteriales which represented 57.3% and 76.6% of the isolates obtained from feline and canine samples, respectively. E. coli was the most frequently isolated bacterium; 54.7% in cats and 55.6% in dogs (Table 3), followed by Proteus mirabilis in dog samples and Enterococcus spp. in cat samples. Additionally, almost 10% of isolates from cats yielded coagulasenegative staphylococci. SIG was the most frequently isolated coagulase-positive Staphylococcus with 4.4% and 7% of isolates from cat and dog samples, respectively.

Antibiotic susceptibility
The great majority of isolates obtained from cat and dog samples (96.7% and 96.1%, respectively) were susceptible to at least one of the oral antibiotics commonly prescribed to UTI patients (ampicillin, amoxicillin-clavulanic acid, cephalothin, tetracycline and trimethoprim-sulfamethoxazole). Out of these, trimethoprimsulfamethoxazole and amoxicillin-clavulanic acid were the drugs with the lowest percentage of resistance (Table 4). Resistance above 20% was detected to cephalothin and cefovecin for E. coli, with the highest resistance against cephalothin (64.4%). Overall, 36.9% of E. coli isolates from canine samples and 39.4% from feline samples were resistant to ampicillin and almost 20% were resistant to amoxicillin-clavulanic acid in cats and dogs. Resistance to erythromycin and tetracycline was detected in approximately 90% and 50% of enterococcal isolates respectively, and 10% of the canine enterococci were resistant to ampicillin and amoxicillin-clavulanic acid, whereas no resistance was observed among the feline enterococci. Fluoroquinolones showed good in vitro activity against the isolates included in our study (resistance 0-7.3% for enrofloxacin and marbofloxacin) excluding coagulase-negative staphylococci with 0% to 17.6% resistance. Staphylococci isolates displayed high resistance to penicillin and ampicillin, but resistance to trimethoprim-sulfamethoxazole and cefovecin was much lower (Table 4).

DISCUSSION
Inadequate empirical choices or antibiotic treatment of nonbacterial conditions such as feline idiopathic/interstitial cystitis can lead to poor patient outcomes and contribute to the selection of resistance, particularly given the long courses of antibiotic treatment usually preferred in companion animal medicine compared to human treatment regimes (Rampacci et al. 2018). Previous  British Small Animal Veterinary Association veterinary guidelines (Weese et al. 2011) advised 7-to 14-day antimicrobial courses but more recent studies indicate that the use of short duration therapy, as used in humans, may be as effective although different dosing regimes and antimicrobials still need to be more specifically evaluated (Weese et al. 2019). Empirical antibiotic choices should take into consideration the local susceptibilities of the uropathogens most likely to be implicated. It is, therefore, important to monitor changes in the aetiology and antibiotic resistance of UTIs in cats and dogs. We detected bacteriuria in 30% (1547/5170) of the urinary samples submitted to the laboratories which was similar to previous reports (Guardabassi et al. 2015, KuKanich & Lubbers 2015. A low prevalence of samples containing two or more isolates was detected as most samples yielded the pure growth of a single organism; this is in agreement with previous studies (Ling et al. 2001, Ball et al. 2008, Lund et al. 2015, Rampacci et al. 2018. The number of urine samples from canine patients submitted for microbiological examination during the span of our study was double that of feline patients which may reflect that UTIs are more common in dogs (Buffington et al. 1997). Another possibility is that sampling is performed more readily in dogs than cats or the higher number of pet dogs than cats in the UK (26% and 31% of the UK households own cats and dogs) could have contributed to this difference (Murray et al. 2010). Positive samples were more commonly from older female dogs, which are known to be at greater risk of developing UTI (Bush 1976, Kivisto et al. 1977, Thomsen et al. 1986, Ling et al. 2001). The mean age of patients with significant bacteriuria, regardless of sex, was approximately 7 years in dogs and 11 years in cats, in agreement with previous studies (Lekcharoensuk et al. 2001, Ling et al. 2001, Cohn et al. 2003.
Amoxicillin, amoxicillin-clavulanic acid, cephalexin, doxycycline and trimethoprim-sulfamethoxazole are defined as oral antibiotics commonly used in the treatment of uncomplicated UTIs (sporadic UTIs in an otherwise healthy animal with a normal urinary tract anatomy and function) (Norris et al. 2000, De Briyne et al. 2014, BSAVA 2018. Fluoroquinolones and second or third-generation cephalosporins are considered second-line antibiotics and are to be used only where there is culture evidence that first-line drugs will not be effective (Norris et al. 2000, Olin & Bartges 2015. E. coli has been recognised as the most common bacterial cause of UTIs in dogs and cats as noted in our study (Ling et al. 2001, Penna et al. 2010, Hall et al. 2013, Dorsch et al. 2015, Moyaert et al. 2017. In agreement with similar studies, resistance above 20% was detected to cephalosporins by E. coli in our study, with the highest (64.4%) being against a first-generation cephalosporin (cephalothin) and 31.2% resistant to a third-generation cephalosporin (cefovecin) in canine E. coli (Weese et al. 2011, Rampacci et al. 2018. This indicates the need to continue monitoring resistance against cephalosporins, especially those of the third generation (e.g. cefovecin) and reserving them solely for difficult cases where culture and susceptibility indicate first-line drugs are not susceptible in order to preserve their efficacy for human and veterinary medicine.
Overall, we observed a predominance of Gram-negative species over Gram-positive species, even though this difference was less marked in samples from feline patients due to a higher prevalence of Enterococcus spp. and coagulase-negative staphylococci. Similarly to previous reports (Litster et al. 2007), Enterococcus spp. was the second most frequent urinary pathogen in cat samples. In humans, this opportunistic organism causes difficult-to-treat infections that usually affect geriatric patients with concurrent diseases and is often associated with implanted medical devices. Furthermore, intrinsic resistance of certain Enterococcus spp. such as Enterococcus faecalis to streptogramins and Enterococcus faecium to β-lactam antibiotics have been previous reported (Fontana et al. 1996, Jackson et al. 2009). In our study, 60% of enterococcal isolates were from samples from patients above the age of 9 years. We detected resistance to erythromycin and tetracycline in approximately 90% and 50% of enterococcal isolates respectively from dogs and cats and although licensed, these antimicrobials are not recommended for UTIs in dogs and cats (BSAVA 2018, Weese et al. 2019). In addition, in vitro resistance does not necessarily correlate with bacteriological cure as most in vitro susceptibility breakpoints are based on achievable serum concentrations, which can be significantly lower than those attained in urine for antibiotics with renal excretion (such as amoxicillin) and in animals that can concentrate urine. In our study, almost 10% of samples from cats with significant bacteriuria yielded the pure growth of coagulase-negative staphylococci. The role of these organisms in canine and feline UTIs is not wholly understood but our findings give support to the view that they are relevant aetiologic agents of these infections. A previous study of urine samples from dogs with UTI implicated Staphylococcus epidermidis, Staphylococcus simulans, Staphylococcus schleiferi and Staphylococcus saprophyticus (Penna et al. 2010) and a more recent study has elucidated the species of coagulase-negative staphylococci recovered from the urinary tracts of both cats and dogs (Moyaert et al. 2017). SIG including the major representative Staphylococcus pseudintermedius are the leading cause of staphylococcal infections in dogs, including UTIs (Lilenbaum et al. 2000, Ganière et al. 2005. It was the most frequently isolated coagulase-positive staphylococci in our study and represented 4.4% and 7.0% of isolates from cat and dog samples, respectively. Amoxicillin is recommended as a first-line choice for the treatment of uncomplicated UTI (Weese et al. 2019). In our study, ampicillin was used to predict susceptibilities to amoxicillin, the latter is preferred for treatment as it has a better oral bioavailability. A considerable number of isolates, particularly E. coli, S. pseudintermedius and coagulase-negative staphylococci, obtained from samples with significant bacteriuria were refractory to ampicillin. Overall, 36.9% of E. coli isolates from canine samples and 39.4% from feline samples were resistant to ampicillin. It is usually recommended that an antibiotic should not be used in the empirical treatment of UTIs if resistance rates exceed 10% to 20% among the most common uropathogens due to an increased risk of poor patient outcomes (Steen 2011). Our results suggest that ampicillin might not be a good antibiotic choice for the empirical treatment of canine and feline UTIs in the UK; however, they should be interpreted with caution. Studies can overestimate the prevalence of antibiotic resistance because they are often biased towards more serious cases (e.g. recurrent UTIs, patients that do not respond to empirical therapy) which are also more likely to be antibiotic resistant (Steen 2011). A lower prevalence of E. coli resistance to ampicillin (approximately 20%) was observed in a recent pan-European study of UTI isolates from cats and dogs not recently exposed to antibiotics (Moyaert et al. 2017).
In our study, clavulanate potentiated amoxicillin was active in vitro against approximately 50% of isolates that were resistant to ampicillin alone. We detected 16.8% to 19.2% prevalence of amoxicillin-clavulanic acid resistance among E. coli isolates, in agreement with previous reports (Marques et al. 2016). Our findings point to the need to monitor the levels of amoxicillin/amoxicillin-clavulanic acid resistance in UTIs of cats and dogs. More recent studies indicate that there is limited evidence for the need to use clavulanic acid as an addition to amoxicillin and its use may not be essential, even with beta-lactamase-producing bacteria, due to the high concentrations of amoxicillin reached in urine (Weese et al. 2019). Trimethoprim-sulfonamides are another advocated first-line option for the empirical treatment of uncomplicated UTIs in dogs and cats (Weese et al. 2011, Rampacci et al. 2018 and, according to our results, might be superior to amoxicillin due to lower resistance rates. Although this should be interpreted with caution due to our low cat sample size (171), this antibiotic has had controversial activity against this bacterial species in other studies with relatively greater associated side effects (Weese et al. 2019).
It is generally recommended that, in order to preserve their effectiveness, the use of broad-spectrum antibiotics should be reserved for cases that do not respond to narrow-spectrum antibiotics and guided by documented in vitro susceptibility of the offending organisms (Weese et al. 2011). Enrofloxacin and marbofloxacin were the antibiotics routinely tested that showed greater in vitro activity against the most common uropathogens (overall susceptibility above 90%) despite previous reports showing an increase in resistance to fluoroquinolones (Cooke et al. 2002, Cohn et al. 2003 and reports after the conduction of this study showing similar results (Maluping et al. 2014, Marques et al. 2016, Rampacci et al. 2018. Research has been pointing to the efficacy of a high-dose, short-course regimen of enrofloxacin in attaining bacteriological cure in uncomplicated UTI cases in dogs and similar results have been obtained with trimethoprim-sulfamethoxazole (Westropp et al. 2012, Olin & Bartges 2015. Short courses of antibiotics have the advantages of increasing owner compliance and reducing the costs and toxicity associated with treatment. Although fluoroquinolones can be effective in the treatment of UTIs, their use in animals is discouraged due to public health and antimicrobial resistance concerns (Weese et al. 2019).
In this study, we used the methods that were the standard of practice at the time of the study, including identification methods and CLSI breakpoints due to their specificity for bacteria isolated from animals (CLSI 2008). Since then new standards have been developed (CLSI VET01S 2020) which raise laboratory testing quality and should be used for previous and current studies in order to obtain the correct susceptibility data in all studies, but this process is far from complete. For example, there are far more breakpoints for pathogens from dogs than those from cats and we lack breakpoints on certain important bacteria which highlights the need to continuously update these standards (Moyaert et al. 2017). In addition, the fact that information on the method of urine collection was often not provided by veterinary surgeons may have further impacted the results as different cut-off for significance are used depending on the collection method. The cutoff for this study was ≥10 5 CFU/mL but significant pathological infections in cystocentesis samples may not have been included as they can be detected at <10 5 and >10 3 CFU/mL (Sørensen et al. 2016).
E. coli continues to be the most common cause of UTIs in dogs and cats with the detection of notable resistance to cephalosporins in our study. Enterococcus spp. were frequently detected in cat samples with high resistance to erythromycin and tetracycline and low resistance to ampicillin and amoxicillin-clavulanic acid. Our findings support the view that Staphylococcus spp. are relevant aetiologic agents of bacterial UTIs indicating the need for further studies on the species of coagulase-negative staphylococci associated with these infections. Considerable resistance to ampicillin and amoxicillin-clavulanic acid from several isolates was identified British Small Animal Veterinary Association which is in contrast to that of trimethoprim-sulfamethoxazole which was low. Enrofloxacin and marbofloxacin showed the best in vitro activity against the isolates tested. Our findings provide evidence of the emergence of resistance to antibiotics commonly used in the treatment of UTIs in dogs and cats among the uropathogens most likely to be responsible for these infections in the UK. Changes in the aetiology and antibiotic resistance of UTIs should be continuously monitored to evaluate the need to introduce changes in the recommendations (e.g. ISCAID guidelines, Weese et al. 2019) about empiric therapy when antimicrobial susceptibility testing is not possible or pending.