Materials and Methods

Dogs—Approximately 1,150 dogs admitted to the Ontario Veterinary College teaching hospital ICU from January 1, 2003, to December 31, 2003, were eligible for inclusion. Dogs in which a urinary catheter was placed at admission or any other time during hospitalization in the ICU were enrolled in the study. Urine samples were collected from all dogs within 24 hours of catheter placement and every 72 hours thereafter. Rectal swab specimens were also collected from dogs with indwelling urinary catheters and from all other dogs admitted to the ICU (approx 985 dogs). Swab specimens were collected within 24 hours of admission and every 72 hours thereafter. Dogs were excluded if underlying illness (eg, severe respiratory distress) made them too unstable for rectal swab specimens to be collected or if the duration of hospitalization in the ICU was < 24 hours. The study was approved by the University of Guelph Animal Care Committee.

Data collected included signalment, cage assignment in the ICU (maximum capacity, 30 cages), dates of entry and discharge from the ICU, tentative and confirmed diagnoses, treatments administered (including antimicrobial agents) during the ICU stay, and duration of catheterization.

Rectal swab specimens, urine samples, and environmental swab specimens—Rectal swab specimens were collected by inserting a cotton-tipped culture swab^a 1 to 2 cm into the rectum and placing in transport tubes. Swab specimens were stored at 4°C (39.2°F), and their contents were suspended in 1 mL of brain-heart infusion broth^b with 20% glycerol in cryovials and stored at −70°C (–94°F) within 24 hours of collection.

All urinary catheters were placed by a registered veterinary technician or a clinician or by a senior veterinary student under the supervision of one of these individuals. Catheters were all made of similar materials. Foley catheters were placed in female dogs, and clear, soft, flexible catheters were placed in male dogs. Approximately 1 mL of urine was aseptically collected from a designated port of a closed collection system attached to the urinary catheter and aseptically transferred via syringe into a sterile tube^c containing no preservative. In most instances, urine samples were stored at 4°C and underwent bacterial culture < 6 hours after collection. Urine samples that were stored at 4°C for a maximum of 24 hours prior to culture were collected in sterile tubes containing boric acid preservative.^d

Every 4 months, environmental samples were collected in the ICU for surveillance studies. Cotton-tipped culture swabs^a were used to sample the surfaces of sinks, cages (bottoms and doors), runs, and drains after cleaning and prior to animal contact. These sites were routinely cleaned with a quaternary ammonium compound^e after animal contact. Clipper blades were cleaned daily and were disinfected with a quaternary ammonium compound^e after clipping wounds and areas of infection. Clipper blades were further cleaned with chlorhexidine disinfectant and swabbed for bacterial culture after cleaning. Additional sites such as the telephone, crash cart, computer keyboards, and door handles, which were not routinely disinfected, were also sampled. Those specimens were submitted to the Animal Health Laboratory at the University of Guelph for culture and susceptibility testing.

Identification of E coli isolates—Rectal swab specimens were obtained from 13 other dogs that were hospitalized in the ICU within 1 to 2 days of the 2 dogs from which MDR E coli was isolated from urine and that were in cages or runs on the same side of the ICU as those dogs. Specimens were plated on MacConkey agar.^f Inoculated plates were incubated at 37°C (98.6°F) and examined for bacterial growth 24 and 48 hours later. Pure cultures of each of the last 5 coliform colonies at the end of the streak were obtained from each plate on days 0, 3, 6, and 9 and used for further testing. A 10-μL sample of urine was plated on MacConkey agar and incubated at 37°C for 24 to 48 hours. Environmental swab specimens were handled in a similar manner at the Animal Health Laboratory.

Colonization was defined as bacteria being cultured from the urinary catheter tip but not the urine. Contamination was defined as cultures from urine yielding a mixed population of bacteria without substantial numbers of a predominant pathogen. Bacterial growth obtained from urinary catheter tips submitted to the Animal Health Laboratory was considered representative of colonization and not infection; those samples were included as environmental cultures with positive results for bacterial growth in the epidemiologic part of the study only. Urinary tract infections diagnosed within 24 hours of admission to the ICU on the basis of bacterial growth from an indwelling urinary catheter were considered community-acquired infections.

Escherichia coli isolates were identified by use of biochemical tests on pure colonies that were derived from urine. Lactose-fermenting and non–lactose-fermenting isolates were tentatively identified as E coli if they were indole positive and oxidase, citrate, urease, and hydrogen sulfide negative and had an acid slope and acid butt in triple sugar iron medium.¹⁶ If reactions to any of those tests were equivocal, an additional panel of tests was performed with a commercial biochemical system^g to confirm the identity of atypical E coli isolates. Subsequently, pure E coli cultures grown from a single colony on MacConkey agar were suspended in brainheart infusion broth with 20% glycerol and frozen at −70°C for future testing. Multi–drug-resistant E coli were defined as urinary or rectal isolates that were resistant to ≥ 3 of the antimicrobial classes tested. Rectal and environmental samples underwent the same biochemical testing for identification of E coli isolates.

Bacterial growth from urine samples was quantified as 1+, 2+, 3+, or 4+ to correlate with growth in 1, 2, 3, or 4 quadrants, respectively, on the agar plates. The number of CFUs/mL of urine was calculated by counting E coli colonies on the plate and multiplying by 10². Mixed infections with E coli were included in the study when > 10² CFUs/mL of the bacteria were recovered.

Antimicrobial susceptibility testing—Susceptibility testing of rectal and urinary E coli isolates was performed via the disk diffusion method according to Clinical and Laboratory Standards Institute guidelines.^17,18 Antimicrobial disks^h used were ampicillin (10 μg), amoxicillin-clavulanate (30 μg), cephalothin (30 μg), cefoxitin (30 μg), cefotaxime (30 μg), ceftiofur (30 μg), trimethoprim-sulfamethoxazole (30 μg), chloramphenicol (30 μg), gentamicin (10 μg), tetracycline (30 μg), enrofloxacin (5 μg), and nalidixic acid (30 μg). Inhibition zone diameters were recorded and interpreted according to Clinical and Laboratory Standards Institute guidelines.^17,18 Urinary isolates of E coli that were resistant to ≥ 3 antimicrobial classes were considered to be MDR and were tested with imipenem(10 μg), carbenicillin (100 μg), and amikacin (30 μg) disks. A reference E coli isolate (American Type Culture Collection 25922) was used with each batch of tests for the rectal and urinary E coli isolates for quality control.^17,18

MDR E coli strain selection for PFGE—Pulsed-field gel electrophoresis was performed on the MDR urinary and rectal E coli isolates from 1 dog. In a second dog, an MDR E coli isolate was not isolated from a rectal swab specimen but was isolated from urine. Rectal swab specimens from 8 dogs in close proximity to that dog that were hospitalized on the same day and from 5 dogs housed in the ICU 1 to 2 days prior to the date of entry of that dog were tested on selective medium for MDR rectal E coli isolates.

PFGE—The method used was adapted from the protocol for E coli O157:H7 described by PulseNet, the National Molecular Subtyping Network for Foodborne Disease Surveillance.¹⁹ Briefly, 5 colonies from a pure culture of each E coli strain on blood agar were mixed in cell suspension buffer (100mM Trisⁱ:100mM EDTA^j [pH, 8.0]) and adjusted to an absorbance of 1.35 to 1.45 at a wavelength of 610 nm. Agarose plugs were made with a final concentration of 1% agarose^k in Tris-EDTA buffer (10mM Tris:1mM EDTA [pH, 8.0]) and proteinase K (20 mg/mL). Each plug was immersed in 1.5 mL of cell lysis buffer–proteinase K mixture (cell lysis buffer, 50mM Tris:50mM EDTA [pH, 8.0] and 1% sarcosine^l; proteinase K, 0.5 mg/mL) and incubated at 54°C (129.2°F) for 2 to 4 hours.

Plugs were washed twice for 15 minutes in 10 mL of sterile water at 50°C (122°F). Three additional washes were performed with Tris-EDTA buffer at 50°C.

Agarose plugs were equilibrated for 15 minutes in the restriction buffer solution. Digestion was performed at 37°C for 4 hours in fresh buffer with 30 U of XbaI^m/plug. Electrophoresis was performed in 1% agarose gels made in 0.5X Tris-borate EDTAⁿ buffer in a pulsed-field electrophoresis apparatus^o at 6.0 V/cm for 15.5 hours with an angle of 120°. Initial and final pulse times were 1 second and 40 seconds, respectively. Thiourea^p (200μM) was added to the gel and electrophoresis buffer when needed to ensure minimal DNA degradation.²⁰ The gels were stained in ethidium bromide for 1 hour, destained in distilled water for 2 hours, and photographed under UV light.

Results

Indwelling urinary catheters were placed in 137 dogs (91 [66%] males and 46 [34%] females) during ICU hospitalization in the 1-year study period. Of those 137 dogs, UTIs were diagnosed in 11 of the 91 (12%) male and 15 of the 46 (33%) female dogs. The overall frequency of UTIs in dogs with indwelling urinary catheters was 26/137 (19%), and the frequency of MDR UTIs was 2/137 (1%).

Community-acquired UTIs—Fifteen dogs had bacterial growth in urine samples obtained within 24 hours of admission to the ICU; in 6 of those 15 dogs, infection was caused by E coli, and of those 6 E coli isolates, 1 isolate was MDR. One dog was excluded from the infection rate calculation because the E coli isolate came from the urinary catheter tip and not the urine; catheter tips are frequently contaminated during removal and do not reflect bacterial populations in the urine.²¹ However, that isolate was used in the epidemiologic part of the study because it represented one of the E coli isolates recovered from dogs in the ICU and was a potential source of infection to other dogs. Staphylococcus aureus was isolated from 3 dogs, and Staphylococcus intermedius, Streptococcus canis, Streptococcus spp, Enterococcus spp, Pseudomonas aeruginosa, and Enterobacter aerogenes were isolated from 1 dog each. Five of the 15 dogs, including the dog with the MDR E coli, had received prior treatment with 1 or more antimicrobial agents.

Catheter-associated UTIs—Fourteen dogs (9 males and 5 females) had long-term (≥ 3 days) urinary catheters placed and urine samples submitted for bacterial culture on days 0 and 3; 3 of those dogs had urine bacterial cultures on days 0, 3, and 6; and 3 had urine bacterial cultures on days 0, 3, 6, and 9. Eleven of the 14 (79%) dogs had an ICU-acquired UTI; 1 of those dogs had bacterial growth from urine samples from days 6 and 9 (Table 1). The frequencies of urine bacterial cultures with positive results for bacterial growth were 9/14 on day 3, 2/3 on day 6, and 1/3 on day 9. In all, there were 6 ICU-acquired E coli isolates. Escherichia coli was cultured from the urinary catheter tip in 1 dog but excluded from infection rate calculations for reasons previously stated. Another E coli isolate was identified on day 6 from another dog that was transferred from the ICU to a general ward with an indwelling urinary catheter on day 2 after catheterization; in that dog, infection was not considered to have been acquired in the ICU. These 2 instances were considered to represent environmental isolates and were not included in calculation of the ICU infection rate.

Of the 6 dogs with ICU-acquired E coli UTIs, 1 was female and 5 were male. Three of the 6 dogs were treated with corticosteroids. One dog received 1 dose of dexamethasone (0.25 mg/kg [0.11 mg/lb], IV) shortly after admission, and the other 2 dogs received dexamethasone (0.10 to 0.25 mg/kg [0.05 to 0.11 mg/lb], IV, q 24 h) for treatment of immune-mediated hemolytic anemia and paresis in the hindquarters. Three dogs had received antimicrobial treatment prior to admission or were being treated with antimicrobials at the time of diagnosis of E coli UTI.

Of the 11 dogs with acquired infections, 4 had mixed infections, including 1 dog with an ICU-acquired E coli UTI. The dog from which E coli was isolated from the catheter tip also had Enterococcus spp isolated from the urine. Another dog had an E coli UTI diagnosed at admission and eliminated by day 3 after admission. However, a newly acquired P aeruginosa UTI was diagnosed in that dog on day 3.

The frequencies of isolation of various bacterial isolates that were in urine at admission and those that were acquired after admission to the ICU were summarized (Table 2). Among the community-acquired infections,E coli was recovered with the highest frequency (40%), followed by S aureus (20%). Among ICUacquired infections, E coli was recovered with the highest frequency (50%), followed by P aeruginosa (33%).

Antimicrobial susceptibility testing—Resistance patterns of E coli isolates detected at admission (day 1) and isolates from ICU-acquired E coli UTIs were summarized (Table 3). The antimicrobials to which the highest number of E coli urinary isolates was resistant were ampicillin and cephalothin. The MDR E coli isolate that was detected at admission was resistant to more antimicrobials than the ICU-acquired MDR E coli. The MDR rectal E coli isolates recovered from the dog with an MDR ICU-acquired E coli UTI had the same antimicrobial resistance pattern as the MDR urinary E coli isolate on days 3, 6, and 9 of hospitalization.

No E coli were recovered from the 60 environmental samples collected over the study period, but other organisms, such as Kluyvera spp and other members of Enterobacteriaceae, were recovered. These were not reported because they were mostly uncommon organisms and not pertinent to the objectives of the study.

Resistance patterns of all urinary E coli isolates— Resistance of E coli to ampicillin and cephalothin was more frequent than resistance to all other antimicrobials (Table 3). Two of the 6 dogs with ICU-acquired E coli UTIs were being treated with ampicillin and 3 were being treated with cefazolin at the time infection was diagnosed. Resistance to cephalothin was detected in 3 E coli isolates, 2 of which were from dogs that had previously been treated with cefazolin.

MDR urinary E coli isolates and PFGE—Both dogs with MDR E coli isolated from urine had been admitted to the ICU from other veterinary hospitals. The dog that had an MDR infection at the time of admission had had previous episodes of UTIs (diagnosed at another veterinary hospital on the basis of urine bacterial culture results or clinical signs) and had received various antimicrobials, including cephalexin, orbifloxacin, and amoxicillin-clavulanate. While in the ICU, that dog was treated with IV administered cefazolin initially and then with ceftiofur after susceptibility results for the MDR E coli isolate were obtained. No E coli were isolated from that dog's rectal swab specimen despite repeated attempts at culturing the organism with enrichment and selective medium. Although the MDR isolate from that dog was collected on day 1, rectal swabs from dogs housed in close proximity to and in the same run prior to admission of the dog were examined as part of the epidemiologic investigation. A dog housed in the same run 1 day previously had 2 rectal MDR E coli isolates with susceptibility patterns that were different from that of the MDR urinary E coli isolate. Both of those rectal isolates were susceptible to enrofloxacin, nalidixic acid, and trimethoprim-sulfamethoxazole, whereas the MDR urinary isolate was resistant to all 3 of those antimicrobials. One of the 2 rectal MDR isolates was resistant to cefoxitin and ceftiofur and had intermediate susceptibility to cefotaxime. The other MDR rectal E coli isolate and the MDR urinary isolate were both susceptible to all 3 extended-spectrum cephalosporins tested. Pulsedfield gel electrophoresis was performed on all 3 isolates, and no genetic similarities were detected among the isolates on the basis of the band patterns.

The MDR E coli urinary isolate obtained from the second dog on days 6 and 9 was resistant to nalidixic acid, enrofloxacin, cephalothin, ampicillin, and chloramphenicol. Antimicrobials administered to that dog while it was hospitalized in the ICU included cefazolin, enrofloxacin, and ampicillin. Multi–drug-resistant E coli was also isolated from rectal swab specimens on days 3, 6, and 9, but not on day 1. Pulsed-field gel electrophoresis was performed to compare the rectal E coli isolates obtained on day 1 with the MDR rectal E coli isolates obtained on days 3, 6, and 9 and the MDR urinaryE coli isolates obtained on days 6 and 9. Three band differences were detected among the MDR urinary isolate from day 6 and the MDR rectal isolates from days 3, 6, and 9, which indicated close genetic relatedness¹² of the urinary and rectal isolates (Figure 1).

Pulsed-field gel electrophoretogram of an MDR urinary Escherichia coli isolate and fecal isolates from a dog with an indwelling urinary catheter and housed in an ICU. M = Molecular weight markers, labeled in kilobases on the left side of the diagram. U = MDR urinary E coli isolate. D0, D6, and D9 = MDR rectal E coli isolates from the same dog on days 0, 6, and 9, respectively;. D3 was not included because the PFGE band pattern was identical to that for D6 and D9. White arrows indicate 3 band differences between the pattern of the urinary tract isolate and the patterns of the D6 and D9 rectal isolates.

Citation: Journal of the American Veterinary Medical Association 229, 10; 10.2460/javma.229.10.1584

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Pulsed-field gel electrophoretogram of an MDR urinary Escherichia coli isolate and fecal isolates from a dog with an indwelling urinary catheter and housed in an ICU. M = Molecular weight markers, labeled in kilobases on the left side of the diagram. U = MDR urinary E coli isolate. D0, D6, and D9 = MDR rectal E coli isolates from the same dog on days 0, 6, and 9, respectively;. D3 was not included because the PFGE band pattern was identical to that for D6 and D9. White arrows indicate 3 band differences between the pattern of the urinary tract isolate and the patterns of the D6 and D9 rectal isolates.

Citation: Journal of the American Veterinary Medical Association 229, 10; 10.2460/javma.229.10.1584

• Download Figure
• Download figure as PowerPoint slide

Pulsed-field gel electrophoretogram of an MDR urinary Escherichia coli isolate and fecal isolates from a dog with an indwelling urinary catheter and housed in an ICU. M = Molecular weight markers, labeled in kilobases on the left side of the diagram. U = MDR urinary E coli isolate. D0, D6, and D9 = MDR rectal E coli isolates from the same dog on days 0, 6, and 9, respectively;. D3 was not included because the PFGE band pattern was identical to that for D6 and D9. White arrows indicate 3 band differences between the pattern of the urinary tract isolate and the patterns of the D6 and D9 rectal isolates.

Citation: Journal of the American Veterinary Medical Association 229, 10; 10.2460/javma.229.10.1584

• Download Figure
• Download figure as PowerPoint slide

Discussion

The 19% frequency of dogs with catheter-associated UTIs in this study was high, compared with findings from other reports.^21–24 However, fundamental differences between the earlier studies and the present study make comparisons difficult. In 1 study²¹ conducted in dogs in a university ICU, the catheter-associated UTI rate was 10%; however, the duration of catheterization in those dogs was < 3 days. In another study,²² a catheterassociated UTI rate of 20% was reported, but that study²² involved healthy female dogs and excluded dogs that were receiving antimicrobial drugs. In a study²⁴ at the Small Animal Clinic at Colorado State University, approximately 25% of all UTIs in dogs from 1983 to 1984 were hospitalacquired, but the rate of nosocomial UTIs decreased to less than half of that by 1987. That study²⁴ incorporated data from a 5-year period and did not distinguish between dogs with or without an indwelling urinary catheter. However, as we did in the present study, those investigators included dogs that were treated with antimicrobials and that had a longer duration of hospitalization (up to 15 days). The rate of catheter-associated UTIs as determined by bacterial culture of urinary catheter tips from dogs in another university ICU was 38%.²³ The duration of catheterization in that study²³ was 2 to 10 days, which was comparable to the present study, and those authors also reported that dogs with catheter-associated UTIs remained in the ICU for a longer time. In the present study, 5 of the 14 (36%) dogs with acquired UTIs were receiving antimicrobials while hospitalized, compared with 78% of the dogs in the study²³ by Lippert et al.

The most common isolate from urine in the present study was E coli, constituting 50% of ICUacquired isolates. Escherichia coli has previously been reported as the most common bacterial isolate from dogs with UTIs at the Ontario Veterinary College teaching hospital.²⁵ The most common organisms isolated from nosocomial UTIs in other veterinary studies have been Klebsiella pneumoniae,E coli,S aureus, Streptococcus spp, Proteus mirabilis, and P aeruginosa.^21–24,26 The frequencies with which the various bacteria were isolated and the bacteria identified in association with UTIs varied among veterinary institutions. Pseudomonas aeruginosa, an opportunistic pathogen that is commonly isolated from nosocomial UTIs associated with urinary catheters,^3,27 was the second most commonly isolated organism from acquired UTIs in dogs in the present study, a finding that may be partly related to that bacterium's ability to form a biofilm on urinary catheters.²⁸

The frequency of UTIs classified as admission infections was 11% (15/137). Among the isolates cultured from admission samples, E coli was the most common, followed by commensal Staphylococcus spp that are part of the resident skin flora in humans and dogs and are frequently isolated from dogs with communityacquired UTIs.^29–31 In the present study, all 4 UTIs caused by Staphylococcus spp were detected on an admission sample and were likely community acquired.

Two dogs in the present study had MDR E coli isolated from urine. Had a more liberal definition of MDR been used (ie, resistance to ≥ 2 antimicrobial classes), larger numbers and possibly more meaningful results may have been yielded. Because of the low frequency of MDR E coli UTIs, few objective conclusions can be drawn about the dynamics of transmission and origins of infection. One dog had been treated with corticosteroids prior to admission and was nonambulatory. The second dog had recurrent episodes of UTIs and had been treated with antimicrobials prior to admission. These risk factors are significantly associated with resistant E coli UTIs in humans^32,33 and may have played a role in these dogs.

Isolates of E coli from dogs in the present study were most frequently resistant to cephalothin and ampicillin. A similar pattern of resistance was reported by Wise et al,²⁴ who observed that ampicillin and cephalosporins were the most frequent antimicrobials used in the small animal clinic at Colorado State University. A high frequency of resistance to ampicillin (40% to 60%) and cephalothin (approx 70%) in E coli isolates from clinical infections was also reported in a more recent study.³⁴

Amoxicillin-clavulanate was not administered to any dogs in the present study during hospitalization in the ICU. One dog had been treated with that medication prior to admission, and 2 dogs with ICU-acquired E coli catheter-associated UTIs were being treated with ampicillin and cefazolin at the time of diagnosis of the infection. It can be speculated that there is potential for development of amoxicillin-clavulanate resistant E coli with administration of ampicillin or cefazolin because these antimicrobials can coselect for resistance to amoxicillin-clavulanate.³⁵ In a study³⁶ in which uropathogenic strains of E coli isolated from dogs were resistant to amoxicillin-clavulanate, the prevalence of resistance was correlated with overuse of that agent as a first-line choice of treatment for UTIs in dogs in the community. It is possible that low levels of resistance to amoxicillinclavulanate exist among urinaryE coli isolates from dogs in the community and an infection with such an isolate could exist at the time of admission to the ICU.

Resistance to extended-spectrum cephalosporins (eg, cefotaxime, cefoxitin, and ceftiofur) or broadspectrum carbapenems (eg, imipenem) was not detected. In the ICU at the veterinary teaching hospital, extended-spectrum cephalosporins and carbapenems are reserved for use in critically ill dogs and cats with a known site of sepsis or are administered on the basis of culture and susceptibility results.

The E coli urinary isolates from the dogs in this study had only a low frequency of resistance to chloramphenicol and tetracycline, which were not administered to any of the dogs either prior to or after admission to the ICU. Bacterial resistance genes for chloramphenicol and tetracycline are frequently linked with resistance to other antimicrobials and may exist on a common plasmid for ampicillin resistance in E coli.¹¹

Reports in the veterinary literature suggest that the increasing use of fluoroquinolones is associated with increasing resistance among E coli isolates.^37–39 In the present study, MDR E coli urinary isolates from 2 dogs were resistant to enrofloxacin; however, only 1 of those dogs was treated with enrofloxacin while in the ICU. Interestingly, the isolate from that dog was detected on day 6, after 3 days of treatment with enrofloxacin. It is possible that treatment with enrofloxacin altered the dog's intestinal flora and selected for MDR E coli organisms, which were subsequently shed in the feces.

The MDR rectal and urinary E coli isolates from 1 dog with a catheter-associated UTI had 3 differences in bands on PFGE, a finding that could be caused by a single genetic event.^12,15 Therefore, the MDR rectal E coli isolates obtained on days 3, 6, and 9 and the MDR urinary E coli isolates obtained from days 6 and 9 from that dog were likely epidemiologically related. Because no MDR urinary or rectal E coli were isolated prior to the third day in the ICU, the MDR urinary E coli isolate likely originated from the dog's intestinal flora. There is another likely explanation why MDR E coli was not cultured from this dog's rectal swab specimen on day 1. The MDR E coli could have originated from another dog and been transferred by a caregiver to this dog, where it colonized the gastrointestinal tract and subsequently caused a catheter-associated UTI. However, no rectal E coli isolates were recovered from the dog in which an MDR urinary E coli isolate was detected at admission. Because that dog was hospitalized for < 3 days, no follow-up rectal swab was obtained. Although that dog was the only dog in this study in which an MDR E coli UTI was diagnosed on admission, the finding confirms that MDR E coli UTIs may be present at admission and may be as important a problem in community-acquired UTIs as in nosocomial UTIs. The only recommendation that can be made on the basis of findings in that dog is that urinary bacterial culture should be performed within 24 hours of urinary catheter placement to determine whether infection existed at admission or developed after placement of the urinary catheter.

Despite the limited population size in this study, there were several key findings. Detection of an MDR E coli UTI at admission indicated that MDR bacteria may be present at admission and are not necessarily ICUacquired infections, and clinicians should monitor patients vigilantly for development of UTIs. The use of genotyping for more accurate identification of individual strains of MDR E coli enabled detection of similar isolates from urine and rectal swab specimens. This finding confirmed that a potential route of UTI is colonization of the urinary catheter by intestinal flora during a dog's stay in the ICU. The antimicrobials that urinary E coli isolates were most frequently resistant to were ampicillin and cephalothin, suggesting that there may be a relationship between antimicrobial use and frequency of resistance.