Small Animals
Clinical performance of a commercial point-of-care urine culture system for identification of bacteriuria in dogs Anna Uhl dvm Faye A. Hartmann ms Katrina R. Viviano dvm, phd From the Department of Medical Sciences (Uhl, Viviano) and the Clinical Pathology Laboratory, UW Veterinary Care (Hartmann), School of Veterinary Medicine, University of Wisconsin, Madison, WI 53706. Dr. Uhl’s present address is Banfield Pet Hospital, 27019 Pacific Hwy S, Des Moines, WA 98198. Address correspondence to Dr. Viviano (viviano@ svm.vetmed.wisc.edu).
OBJECTIVE To evaluate the clinical performance of a commercially available compartmentalized urine culture and antimicrobial susceptibility test plate (CCSP) for identification of canine bacteriuria and assessment of isolate antimicrobial susceptibility. DESIGN Cross-sectional study. ANIMALS 71 dogs. PROCEDURES Urine samples (n = 84) were divided into 3 aliquots. One aliquot (reference culture) was plated on culture medium ≤ 1 hour after collection for quantitative culture and testing by standard laboratory methods, another was stored at 4°C for 24 hours (to mimic storage practices at primary care facilities) and then processed by standard methods, and the third was applied to a CCSP ≤ 1 hour after collection to be processed and interpreted according to manufacturer instructions. Results were compared with those for reference culture, which was used as the criterion reference standard. Sensitivity, specificity, positive and negative predictive values, and agreement between methods was evaluated. RESULTS 43 isolates (25 single and 9 multiple isolates) were identified in 34 reference cultures. All results for stored cultures were identical to those for reference cultures. Overall sensitivity of the CCSP method to detect bacteriuria was 93%, and specificity was 100%. Thirty-three of 43 (77%) and 19 of 33 (58%) CCSP bacterial isolates were correctly identified to the genus and species level, respectively. The CCSP antimicrobial susceptibility results matched those for reference cultures for 13 of 33 (39%) isolates evaluated. CONCLUSIONS AND CLINICAL RELEVANCE Limitations of the CCSP method included inaccuracy of some antimicrobial susceptibility test results and failure to correctly identify bacteriuria in some dogs. ( J Am Vet Med Assoc 2017;251:922–928)
B
acterial UTIs are common in dogs. Authors of 1 report1 estimated that 14% of dogs have ≥ 1 UTI during their lifetime. A UTI is diagnosed on the basis of clinical signs consistent with the condition and documented bacteriuria. The gold standard for identification of bacteriuria is a positive result for quantitative bacterial culture of urine (≥ 103 CFUs of bacteria/mL) with concurrent antimicrobial susceptibility testing to optimize treatment and minimize development of antimicrobial resistance.2 The immediate plating of urine after collection for bacterial culture is recommended to minimize false-negative or false-positive results.3,4 The practical alternative to immediate culture has been urine storage, either at 4°C without the adABBREVIATIONS CCSP Compartmentalized culture and antimicrobial susceptibility testing plate NPV Negative predictive value PPV Positive predictive value UTI Urinary tract infection 922
dition of a preservative or at room temperature with the addition of boric acid, followed by plating of the sample for culture ≤ 24 hours after collection.5–8 Although urine is a good storage medium, both strategies have limitations, especially for the preservation of Escherichia coli, one of the most common urinary pathogens isolated.3,8 Urine storage may alter bacterial growth, especially if a sample is not refrigerated within 1 to 2 hours after collection.3,8 The use of boric acid as a preservative has been reported to have an inhibitory effect on bacterial growth in urine samples from people7,9 and dogs.10 Unless a veterinary practice has a microbiology laboratory, most veterinarians must store urine samples and then ship them ≤ 24 hours after collection to a commercial laboratory for bacterial culture and antimicrobial susceptibility testing. Anecdotal information from primary care veterinary practices indicates bacterial urine cultures are often not pursued because of the associated cost and the perceived prolonged turnaround time. In addition, the
JAVMA • Vol 251 • No. 8 • October 15, 2017
Small Animals
impression of some practitioners is that there can be discordance between the clinical history or clinical signs of a dog suspected to have a UTI and the results obtained from urine cultured after storage for 24 hours (ie, a presumed false-negative result associated with urine storage). A quantitative in-clinic urine culture that could be performed immediately after sample collection would be an ideal method, but for most practice situations, this has not been practical or economical. Historically, point-of-care urine culture kits have been marketed and used in human medicine for the assessment of bacteriuria.11–13 Recently, similar urine culture kits (eg, a dipstick paddle system14 and CCSPs15,16) have been described for confirmation of bacteriuria and facilitation of bacterial identification in dogs. The CCSPs also provide antimicrobial susceptibility results as a means to aid antimicrobial selection. These systems were designed in part to address the limitations of having to send out urine samples for microbial culture and antimicrobial susceptibility testing by a microbiology laboratory. To the authors’ knowledge, independent field investigations on the clinical performance of these veterinary point-of-care urine culture systems have been limited to 2 studies14,16 published prior to this investigation and 1 study15 that was published after the completion of the study reported here. Those studies compared canine urine culture results obtained through a university microbiology laboratory by standard quantitative methods with results of either the dipstick paddle14 or CCSP15,16 systems. The reported sensitivities and specificities for the CCSP system were 81% to 83% and 99% to 100%, respectively, and those for the dipstick paddle system were 92% and 99%, respectively. The purpose of the study reported here was to evaluate use of the CCSP system to identify bacteriuria and the uropathogens present in a population of dogs at risk for bacteriuria by comparing test results with those obtained by standard microbiological methods. An additional objective was to determine whether antimicrobial susceptibility data obtained with the CCSP system would result in appropriate clinical antimicrobial selection for dogs from which a single uropathogen was isolated. We also sought to determine whether urine storage with refrigeration for 24 hours (to mimic the typical situation in primary care practices) would affect the results of culture by standard microbiological methods.
Materials and Methods Dogs and samples
Urine samples from client-owned dogs evaluated at UW Veterinary Care from June 11, 2013, to December 30, 2013, were used in this prospective, observational study. Only samples from dogs considered to be at risk for bacteriuria on the basis of history
or clinical findings and for which the clinician pursued a microbial culture and antimicrobial susceptibility testing as part of the clinical work-up were included. Dogs of any age, breed, or sex that had urine collected by cystocentesis, catheterization, or clean midstream catch of a voided sample were eligible for inclusion. Dogs were excluded from the study if the sample volume was insufficient (ie, < 5 mL) to ensure adequate amounts for both the clinically requested diagnostics and the tests performed for research purposes. All study protocols were reviewed, approved, and conducted in accordance with the requirements of the University of Wisconsin Animal Care and Use Committee. Data collected and recorded for all dogs enrolled in the study included age, sex, breed, weight, medical history, medications being administered at time of enrollment and antimicrobial treatments given in the previous 7 days, clinical diagnosis, and treatments recommended by the clinician responsible for the case (including antimicrobial administration, when prescribed).
Sample handling
Each urine sample was divided into 3 aliquots and placed into sterile plastic containers without additives after collection. Less than 1 hour after sample collection, the first aliquot was plated on standard urine culture medium for quantitative culture and antimicrobial susceptibility testing by standard laboratory methods for diagnostic purposes as requested by the attending clinician. These methods (with samples plated ≤ 1 hour after collection) were considered the clinical criterion reference standard methods for study purposes. Immediately after plating of this sample (ie, the reference culture sample), the remaining aliquots were processed. One of the aliquots was refrigerated at 4°C for subsequent culture and testing by standard laboratory methods, and the other was applied to a CCSP.a
Reference cultures
Quantitative bacterial culture consisted of inoculating 10 µL of urine collected by clean midstream catch or 100 µL of urine collected by cystocentesis onto both trypticase soy agar supplemented with sheep blood (to a 5% concentration) and eosin–methylene blue agar.b Urine cultures were evaluated for growth after 24 and 48 hours of incubation at 36°C in 5% CO2 atmosphere. Growth was quantitated as the number of CFUs per mL. Bacterial growth cutoffs were defined for each urine collection method used on the basis of recommendations in human and veterinary medical literature.17–21 For the present study, a positive result from a voided sample was defined as 1 to 3 microbial species present at concentrations > 10,000 CFUs/mL. The presence of > 3 species of microorganisms or concentrations < 10,000 CFUs/mL indicated probable urethral contamination, and results of these cultures
JAVMA • Vol 251 • No. 8 • October 15, 2017
923
Small Animals
were considered to represent unremarkable growth. Any growth of microorganisms from urine collected by cystocentesis was considered clinically relevant unless it was suggestive of gastrointestinal content contamination during cystocentesis. All cultures that yielded unremarkable growth were scored as having negative results for purposes of the study. Identification of bacterial isolates was made on the basis of colony type and morphology, Gram-staining characteristics, and standard biochemical test results, and information obtained with a commercial identification system.c Minimal inhibitory concentrations were determined for these isolates by use of a commercial systemc,d according to the manufacturer’s recommendations. Quality control was performed and antimicrobial susceptibility results were interpreted in accordance with Clinical Laboratory Standards Institute guidelines.22–24 Intermediate susceptibility interpretations were considered resistant for purposes of comparison in this study.
Storage of samples prior to culture
The designated urine sample aliquot was refrigerated at 4°C for 24 hours to mimic the standard procedure for storing urine before and during shipment to a microbiology laboratory. After the 24-hour storage time, the urine was cultured by the same standard microbiological methods as described for the reference culture samples. Bacterial quantitation, isolate identification, and antimicrobial susceptibility testing for these samples (deemed stored sample cultures) were performed as described for reference cultures by personnel in the microbiology section of the laboratory.
CCSP testing
The CCSPa used was designed to aid in the diagnosis of UTIs in dogs by providing semiquantitative bacterial counts and antimicrobial susceptibility results for uropathogens isolated from fresh samples or samples stored at 2° to 8°C for < 48 hours.25 All samples used with the CCSPs in the present study were fresh (applied ≤ 1 hour after collection). The plates were used in accordance with the manufacturer’s package insert25 prior to their expiration date. Plates were stored at 2° to 8°C and warmed to room temperature prior to use. Culture plating was performed by 1 investigator (AU) over the course of the study. Each plate consisted of a Petri dish containing ready-to-use Mueller-Hinton II agar divided into 1 large compartment free of antimicrobials (used for colony counts) and 5 smaller compartments, each containing 1 antimicrobial (ampicillin, amoxicillin-clavulanate, cephalothin, enrofloxacin, or trimethoprim-sulfamethoxazole; used for susceptibility testing of microbes).25 Urine samples were plated on CCSPs (1 plate/ sample) according to the manufacturer’s recommendations. Briefly, urine (approx 1 to 2 mL) was flooded onto the plate for 3 to 5 seconds in a manner that ensured all compartments were coated, and any ex924
cess urine was poured off the plate. The plate was then incubated in ambient atmospheric conditions at 35°C and checked for growth by gross visual examination after 24 hours of incubation by the same investigator (AU), who was unaware of the reference culture results. The plate was incubated for an additional 24 hours if no growth was observed at the first inspection. Clinically relevant bacterial growth was determined by visual inspection of the large counting compartment, and the results for each sample were determined according to the interpretive guidelines provided by the manufacturer. Briefly, bacterial isolates were identified by color changes of the medium specific to bacterial species, and an estimate of the number of CFUs was made on the basis of colony density as described and illustrated by the manufacturer’s instructions.25 Total colony counts from voided samples and urine collected by cystocentesis were interpreted in a manner similar to that used for reference culture results. For plates deemed to have clinically relevant bacterial growth, the antimicrobial susceptibility compartments of the plate were evaluated and results were recorded. Growth in each of these compartments was compared with that in the larger antimicrobial-free compartment to determine whether the isolate was susceptible or resistant to the antimicrobials included in the plate (an interpretation of intermediate could not be determined by this method). An isolate was considered susceptible if there was no growth or resistant if there was growth in a particular antimicrobial compartment.
Statistical analysis
Results of a power calculation performed on the basis of data from a study11 investigating use of a similar CCSP-based urine culture system in human patients indicated that 70 urine samples would be needed to achieve 80% power to detect bacteriuria in dogs. The Cohen κ coefficient was calculated (with 95% confidence interval) as a quantitative measure of the magnitude of agreement between culture methods beyond that expected by chance. Positive and negative κ coefficients indicate agreement greater than or less than that expected by chance alone, respectively. The degree of agreement was interpreted as described elsewhere (1 = perfect; 0.80 to 0.99 = very good; 0.6 to 0.79 = good; 0.4 to 0.59 = moderate; 0.2 to 0.39 = fair; and 0 to 0.2 = poor).26 The sensitivity, specificity, PPV, NPV, and accuracy of the CCSP method (with 95% confidence interval) were calculated with standard equations.26 For these indices of clinical performance, 1 sample/dog was included in the analyses. In cases where multiple samples had been obtained from a given dog and cultures were repeated, only the results of the first culture were included in the statistical analysis.
Results Eighty-four urine samples collected by cystocentesis (n = 48), catheterization (4), or clean mid-
JAVMA • Vol 251 • No. 8 • October 15, 2017
Small Animals
stream catch of a voided sample (32) were included in the study. Samples were collected from 71 dogs (25 males [22 neutered and 3 sexually intact] and 46 females [40 spayed and 6 sexually intact]). There were 63 purebreds and 8 mixed-breed dogs; the 5 breeds most frequently represented were Labrador Retriever (n = 7), Beagle (5), Bichon Frise (5), American Pit Bull Terrier (4), and Dachshund (4). The median age of dogs was 8.6 years (range, 1 to 15 years), and median weight was 19 kg (42 lb; range, 4.6 to 96.9 kg [10.1 to 213.2 lb]). Clinical diagnoses included UTI (n = 35), neoplasia (9 [3 and 6 involving and not involving the urinary bladder, respectively]), proteinuria (6), diabetes mellitus (3), acute kidney injury (3), and urinary incontinence (3). Twelve dogs had other clinical diagnoses, including inappropriate urination attributed to behavioral causes, hematuria of undetermined cause, vaginitis, renal dysplasia, chronic kidney disease, hyperadrenocorticism, hypoadrenocorticism, protein-losing nephropathy, nephrolithiasis, pyogranulomatous lymphadenitis, sepsis, and uveitis. At the time of urine culture, 21 dogs were receiving or had received antimicrobial treatment within the past 7 days. These treatments included β-lactams (n = 13), fluoroquinolones (4), trimethoprim-sulfamethoxazole (2), clindamycin (1), and metronidazole (1).
Bacterial isolates
Culture results attained by standard microbiological methods (for samples plated ≤ 1 hour after collection [reference cultures] and those plated after refrigeration for 24 hours [stored sample cultures]) and by use of the CCSP were summarized (Table 1). The results for stored sample cultures were identical to those of the reference cultures for all tests, including detection of a positive result, genus and species
identification of isolates, and results of antimicrobial susceptibility testing. Compared with the 43 isolates identified in reference cultures, the CCSP correctly identified 33 (77%) isolates to the genus level, with 19 of 33 (58%) isolates identified to the species level and 14 of 33 (42%) identified only to the genus level (Table 1). The 33 isolates detected by CCSP included 13 of 19 (68%) gram-positive and 20 of 24 (83%) gram-negative isolates, compared with reference culture results. Isolates identified in reference cultures included E coli (n = 18), Staphylococcus pseudintermedius (7) and methicillin-resistant S pseudintermedius (2), Streptococcus canis (7), Proteus mirabilis (3), Enterobacter cloacae (2), Enterococcus faecium (2), Enterococcus faecalis (1), and Pseudomonas aeruginosa (1). Results for the same samples cultured by use of CCSPs revealed identification of E coli (n = 16), E cloacae (1), E faecalis (1), and P aeruginosa (1). Isolates identified only at the genus level, as allowed for by the CCSP method, included Staphylococcus spp (7), Streptococcus spp (3), Proteus spp (2), and Enterococcus spp (2). Cohen κ analysis confirmed perfect agreement between results (positive vs negative) obtained with reference cultures and stored sample cultures (κ = 1). Agreement between results (positive vs negative) for CCSPs and reference cultures was very good when all dogs (κ = 0.94) and only dogs that had a single bacterial isolate identified in reference cultures (κ = 0.93) were included in the analysis. Agreement for CCSPs and reference cultures was fair (κ = 0.36) for samples from dogs in which multiple bacterial isolates were identified in reference cultures (Table 2). The sensitivity, specificity, PPV, NPV, and accuracy of the CCSP method to detect clinically relevant
Table 1—Summary of results for 84 urine samples cultured and tested by standard laboratory methods (with samples plated ≤ 1 hour after collection [clinical reference culture] or after storage at 4°C for 24 hours [stored sample culture]) and by use of a CCSP system. Variable
Clinical Stored reference culture sample culture
Culture result (No. of samples) Positive 34 34 Single isolate 25 25 Multiple isolates 9 9 Negative 50 50 Bacterial isolate identification No. of isolates 43 43 Gram-positive 19 19 Gram-negative 24 24 Genus and species identified 43 43 Genus identified only 0 0
CCSP culture 30 27 3 54 33 13 20 19 14
Urine was collected from 71 dogs considered to be at risk for bacteriuria. Samples were collected by cystocentesis (n = 48), catheterization (4), or clean midstream catch of a voided sample (32) and divided into 3 aliquots (1/method). A positive culture result was defined as clinically relevant growth; for the reference culture method, cutoffs were determined as described elsewhere,17–21 and for the CCSP, this was determined according to the manufacturer’s instructions for use of the plates. The reference culture result was used for diagnostic purposes for dogs in the study.
JAVMA • Vol 251 • No. 8 • October 15, 2017
925
Small Animals
Table 2—Sensitivity, specificity, PPV, NPV, and accuracy for the use of CCSPs to identify bacteriuria in urine samples from 71 dogs, compared with results for reference cultures and assessment of agreement between the 2 methods as measured by Cohen k analysis. PPV Test
NPV
No. of dogs Sensitivity (%) Specificity (%) Accuracy (%) TP/TP+FP % TN/TN+FN %
Bacteriuria (all dogs) Bacterial isolate identified (genus) from cultures with a single isolate† Bacterial isolate identified (genus) from cultures with multiple isolates‡
κ*
71 63
93 (77–99) 90 (69–98)
100 (92–100) 100 (93–100)
97 (91–100) 97 (91–100)
27/27 19/19
100 (87–100) 100 (82–100)
42/44 42/44
95 (85–99) 0.94 (0.86 to 1) 96 (87–99) 0.93 (0.84 to 1)
50
25 (3.2–65)
100 (92–100)
88 (79–97)
2/2
100 (16–100)
42/47
88 (75–95) 0.36 (–0.01 to 0.73)
Data are shown as value (95% confidence interval) where appropriate. When multiple samples were cultured for a given dog, only the first sample was included in these analyses. *The κ coefficient was used to determine the degree of intermethod agreement as follows: 1= perfect, 0.80 to 0.99 = very good, 0.6 to 0.79 = good, 0.4 to 0.59 = moderate, 0.2 to 0.39 = fair, and 0 to 0.2 = poor. †Eight samples that had multiple isolates identified by the reference culture method were excluded from this analysis. ‡For dogs with multiple isolates, calculations were based on correct versus incorrect identification of all isolates. FN = Number of false-negative results. FP = Number of false-positive results. TN = Number of true-negative results. TP = Number of true-positive results. See Table 1 for remainder of key.
Table 3—Empirical antimicrobial treatments and antimicrobial susceptibility profiles for isolates from urine samples of 9 dogs that had positive culture results for a single isolate and discrepancies in susceptibility data between the reference culture and CCSP methods. Antimicrobial susceptibility Empirical Isolate and treatment Dog No. culture method AMP AMC CEP ENR TMS prescribed 1 2 3* 4* 5* 6 7 8* 9*
E faecalis RC CCSP E coli RC CCSP E coli RC CCSP E coli RC CCSP E coli RC CCSP S pseudintermedius RC CCSP E coli RC CCSP E coli RC CCSP E coli RC CCSP
S S — I — AMX S S R R S S S S S S AMC S S R S S S S I S S AMX R S S S S S S S S S AMX R R R S S S S I S S AMC R R R S S R S S S S AMC S S S S S R S I S S CIP R R R S S S S I S S CEF R R R S S R S S S R AMC R R R S S
Antimicrobial susceptibility testing for reference cultures was performed by standard microbiology methods with results (susceptible [S], intermediate [I], or resistant [R]) interpreted according to Clinical Laboratory Standards Institute guidelines.21–23 Results for testing with CCSPs were interpreted according to the manufacturer’s instructions. *Reliance on CCSP results alone for these dogs could have resulted in an unnecessary change in empirical antimicrobial treatment. — = Not applicable. AMC = Amoxicillin-clavulanate. AMP = Ampicillin. AMX = Amoxicillin. CEF = Cefpodoxime. CEP = Cephalothin. CIP = Ciprofloxacin. ENR = Enrofloxacin. RC = Reference culture. TMS = Trimethoprim-sulfamethoxazole.
bacterial growth in samples from all dogs, dogs with single bacterial isolate identified, and dogs with multiple bacterial isolates identified were summarized (Table 2). One or more isolates were missed by use 926
of the CCSP in 6 of 9 samples from dogs confirmed to have mixed bacterial infections and 4 of 25 samples from dogs with a single isolate identified (including multiple cultures from individual dogs).
JAVMA • Vol 251 • No. 8 • October 15, 2017
Small Animals
Antimicrobial susceptibility
Comparison of antimicrobial susceptibility testing results by standard microbiological methods (for bacterial isolates from reference cultures) versus those obtained with CCSPs yielded the same results for 13 of 33 (39%) isolates that were identified by both methods. For the remaining 20 (61%) isolates, the in vitro antimicrobial susceptibility results by CCSP conflicted with those for the reference method. These disparate results were attributed to identification of different antimicrobial susceptibility interpretations (susceptible vs resistant) for a particular antimicrobial with the 2 methods (n = 18) or inappropriate antimicrobial testing with CCSPs (2; both Enterococcus spp) owing to fixed antimicrobial testing. Susceptibility testing results for E coli isolated from 16 samples by use of CCSPs were incorrect for ≥ 1 of the β-lactam antimicrobials tested (ampicillin [n = 8], amoxicillin-clavulanic acid [8], or cephalothin [4]). Four of 7 Staphylococcus spp isolates were incorrectly interpreted as being susceptible to ampicillin by CCSP, and 1 of 3 isolates of Streptococcus spp was incorrectly identified as resistant to enrofloxacin. When antimicrobial susceptibility test results for samples that had only 1 isolate identified by both methods (n = 21) were compared, findings were the same for 11 (52%) and conflicting for 10 (48%). Disparate results for 8 of 10 isolates were primarily attributed to incorrect interpretation results for susceptibility to β-lactams. Review of these cases revealed that empirical antimicrobial treatment was initiated for 9 dogs with clinically relevant bacteriuria caused by single isolates that also had discrepancies in antimicrobial susceptibility interpretations between the 2 methods. On the basis of results for the CCSP testing alone, empirical treatment could have been unnecessarily changed for 5 of 9 dogs. In comparison, empirical treatment might have been changed for only 1 of these 9 dogs on the basis of reference culture results (Table 3).
Discussion Results of the present study confirmed that storage of canine urine samples at 4°C for 24 hours did not affect the results of culture and susceptibility testing of isolates by standard microbiological methods. Results of bacterial genus and species identification as well as antimicrobial susceptibility testing for isolates from stored sample cultures were identical to those for reference cultures (plated ≤ 1 hour after sample collection for culture and testing by the same methods), supporting that storage of urine overnight at 4°C is a viable clinical alternative when immediate urine culture is not possible. In addition, samples cultured and evaluated on CCSPs had very good agreement with reference cultures for identification of clinically relevant bacterial growth or for single-isolate infections, but only fair agreement for polymicrobial infections, in which isolates were sometimes not identified. On the basis of the calculated overall sensitivity (93%) and specificity (100%) of the CCSP method in this study, growth on the CCSP culture is a good indicator of bacteriuria (PPV, 100%); however, a no-growth CCSP re
sult could miss identifying bacteriuria (NPV, 95%) in this population of dogs at risk for the condition. For dogs with positive bacterial culture results, our findings indicated that the antimicrobial susceptibility data from the CCSP culture should be interpreted cautiously when making therapeutic decisions, especially for the use of β-lactam antimicrobials. In this study, reliance on only the CCSP antimicrobial susceptibility results could have resulted in an unnecessary change or escalation (eg, use of fluoroquinolones) in antimicrobial administration for 5 of 9 dogs. Furthermore, of the 7 Staphylococcus spp isolates identified by both methods, the CCSP falsely indicated susceptibility to ampicillin for 4 isolates and missed detection of methicillin resistance for 2 isolates owing to a lack of oxacillin susceptibility testing by the CCSP plate. In addition, reliance on the CCSP antimicrobial susceptibility results for E faecalis isolated from 1 sample could have led to the inappropriate selection of trimethoprim-sulfamethoxazole for treatment; although an isolate may appear susceptible to the drug in vitro, it is not clinically effective against enterococci.22 These findings suggested that use of antimicrobial susceptibility data obtained by interpretation of CCSP results may not promote improved antimicrobial use and could potentially contribute to treatment failure or increased antimicrobial resistance of uropathogens. The primary advantage of CCSP use as a point-ofcare diagnostic aid is its minimal technical training and equipment requirements, the short set-up time needed after urine collection, and avoidance of the need for storage of urine before it can be cultured. The use of this inclinic method has the potential to be less expensive than shipping urine specimens to a microbiology laboratory for evaluation. Furthermore, plates can be evaluated 24 hours after culture is initiated and might provide some preliminary information to aid in case management, although results must be interpreted in the context of the patient’s clinical history and response to treatment. Some disadvantages of CCSP use are that plating requires a fairly large urine volume (minimum of 1.5 mL); antimicrobial susceptibility testing is limited to 5 antimicrobials, which may be inappropriate for some isolates; and susceptibility test results, as identified in this study and others,15,16 frequently conflict with those obtained by standard microbiological methods. In addition, the interpretation of the results relies on colorimetric changes, dependent on the evaluator’s subjective judgement for isolate identification. In the present study, accurate identification of isolates was subjectively difficult or impossible in some samples, resulting in inconsistencies between results for this method and that used for reference cultures, particularly in some samples with multiple bacterial species present, low bacterial counts, or a color change that was difficult to judge. The results of this study were similar to findings of another clinical study16 that evaluated use of CCSPs for assessment of canine urine samples. In that study, results obtained with the CCSP method were compared with results of quantitative urine culture performed by a university microbiology laboratory in samples plated
JAVMA • Vol 251 • No. 8 • October 15, 2017
927
Small Animals
≤ 24 hours after collection. The reported sensitivity and specificity of the test (evaluated in samples from 96 dogs) was 81% and 99%, respectively. The results of both (previous and present) studies supported that quantitative urine culture with antimicrobial susceptibility testing through a microbiology reference laboratory is a more accurate means to aid in the diagnosis of canine bacteriuria and direct antimicrobial selection. Limitations of the present study included the relatively small number of canine urine specimens of adequate volume for both clinical diagnostic testing and use for research purposes, which limited power of the study. The overall low number of samples that had positive culture results limited comparisons among methods for various bacterial species. Additionally, the underlying conditions identified in this group of dogs were diverse, and dogs with bacteriuria included patients with conditions other than uncomplicated UTIs. However, the diversity of patients included in the study reflects to some degree the clinical population of dogs for which urine culture may be pursued. Finally, the handling of samples that were stored with refrigeration prior to culture might not have adequately represented the conditions of urine storage and transportation in practice. The results of this study of a clinical population of dogs suspected to have bacteriuria did not support the use of the CCSP method as a replacement for standard microbiological methods for bacterial culture and antimicrobial susceptibility testing of isolates from canine urine samples. However, for clinical situations in which a quantitative culture is not available, knowledge of the clinical limitations of the CCSP method is important for clinicians considering a point-of-care bacterial culture method.
5.
6. 7. 8. 9. 10. 11. 12.
13. 14.
15.
16. 17.
Acknowledgments No third-party funding or support was received in connection with this study or the writing or publication of the manuscript. The authors declare that there were no conflicts of interest and issue no disclaimers. Presented in part as a poster at the American College of Veterinary Internal Medicine Forum, Nashville, Tenn, June 2014.
18.
Footnotes
20.
a. b. c. d.
Flexicult Vet, Atlantic Diagnostics, Miami, Fla. Remel, Lenexa, Kan. Vitek 2, BioMérieux, Marcy l’Étoile, France. TREK Diagnostic Systems, Cleveland, Ohio.
References
21. 22.
23.
1. Ling GV. Therapeutic strategies involving antimicrobial treatment of the canine urinary tract. J Am Vet Med Assoc 1984;185:1162–1164. 2. Weese JS, Blondeau JM, Boothe D, et al. Antimicrobial use guidelines for treatment of urinary tract disease in dogs and cats: antimicrobial guidelines working group of the international society for companion animal infectious diseases. Vet Med Int 2011;2011:263768. 3. Hindman R, Tronic B, Bartlett R. Effect of delay on culture of urine. J Clin Microbiol 1976;4:102–103. 4. Perrin J, Nicolet J. Influence of the transport on the outcome of the bacteriological analysis of dog urine compari-
928
19.
24.
25. 26.
son of three transport tubes. Zentralbl Veterinarmed B 1992;39:662–667. Horton JA III, Kirshblum SC, Linsenmeyer TA, et al. Does refrigeration of urine alter culture results in hospitalized patients with neurogenic bladders? J Spinal Cord Med 1998;21:342–347. Lum KT, Meers PD. Boric acid converts urine into an effective bacteriostatic transport medium. J Infect 1989;18:51–58. Meers PD, Chow CK. Bacteriostatic and bactericidal actions of boric acid against bacteria and fungi commonly found in urine. J Clin Pathol 1990;43:484–487. Padilla J, Osborne CA, Ward GE. Effects of storage time and temperature on quantitative culture of canine urine. J Am Vet Med Assoc 1981;178:1077–1081. Gillespie T, Fewster J, Masterton RG. The effect of specimen processing delay on borate urine preservation. J Clin Pathol 1999;52:95–98. Rowlands M, Blackwood L, Mas A, et al. The effect of boric acid on bacterial culture of canine and feline urine. J Small Anim Pract 2011;52:510–514. Blom M, Sorensen TL, Espersen F, et al. Validation of Flexicult SSI-urinary kit for use in the primary health care setting. Scand J Infect Dis 2002;34:430–435. Scarparo C, Piccoli P, Ricordi P, et al. Evaluation of the DipStreak, a new device with an original streaking mechanism for detection, counting, and presumptive identification of urinary tract pathogens. J Clin Microbiol 2002;40:2169–2175. Winkens R, Nelissen-Arets H, Stobberingh E. Validity of the urine dipslide under daily practice conditions. Fam Pract 2003;20:410–412. Ybarra WL, Sykes JE, Wang Y, et al. Performance of a veterinary urine dipstick paddle system for diagnosis and identification of urinary tract infections in dogs and cats. J Am Vet Med Assoc 2014;244:814–819. Guardabassi L, Hedberg S, Jessen LR, et al. Optimization and evaluation of Flexicult(R) Vet for detection, identification and antimicrobial susceptibility testing of bacterial uropathogens in small animal veterinary practice. Acta Vet Scand 2015;57:72. Olin SJ, Bartges JW, Jones RD, et al. Diagnostic accuracy of a point-of-care urine bacteriologic culture test in dogs. J Am Vet Med Assoc 2013;243:1719–1725. Burd EM, Kehl KS. A critical appraisal of the role of the clinical microbiology laboratory in the diagnosis of urinary tract infections. J Clin Microbiol 2011;49:S34–S38. Carter JM, Klausner JS, Osborne CA, et al. Comparison of collection techniques for quantitative urine culture in dogs. J Am Vet Med Assoc 1978;173:296–298. Comer KM, Ling GV. Results of urinalysis and bacterial culture of canine urine obtained by antepubic cystocentesis, catheterization, and the midstream voided methods. J Am Vet Med Assoc 1981;179:891–895. Sanderson SL. Urinary system introduction. In: Kahn C, ed. The Merck veterinary manual. 10th ed. Whitehouse Station, NJ: Merck & Co, 2010;1380. Stamm WE, Wagner KF, Amsel R, et al. Causes of the acute urethral syndrome in women. N Engl J Med 1980;303:409–415. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing. CLSI document M100–23. Wayne, Pa: Clinical and Laboratory Standards Institute, 2013. Clinical and Laboratory Standards Institute. Performance Standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals; approved standard. 4th ed. CLSI document VET01–A4. Wayne, Pa: Clinical and Laboratory Standards Institute, 2013. Clinical and Laboratory Standards Institute. Performance Standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals. CLSI document VET01–S2. Wayne, Pa: Clinical and Laboratory Standards Institute, 2013. Flexicult vet [package insert]. Miami, Fla: Atlantic Diagnostics, 2013. Altman D. Practical statistics for medical research. London: Chapman & Hall, 1997.
JAVMA • Vol 251 • No. 8 • October 15, 2017
Clinical performance of a commercial point-of-care urine culture system for identification of bacteriuria in dogs Anna Uhl dvm Faye A. Hartmann ms Katrina R. Viviano dvm, phd From the Department of Medical Sciences (Uhl, Viviano) and the Clinical Pathology Laboratory, UW Veterinary Care (Hartmann), School of Veterinary Medicine, University of Wisconsin, Madison, WI 53706. Dr. Uhl’s present address is Banfield Pet Hospital, 27019 Pacific Hwy S, Des Moines, WA 98198. Address correspondence to Dr. Viviano (viviano@ svm.vetmed.wisc.edu).
OBJECTIVE To evaluate the clinical performance of a commercially available compartmentalized urine culture and antimicrobial susceptibility test plate (CCSP) for identification of canine bacteriuria and assessment of isolate antimicrobial susceptibility. DESIGN Cross-sectional study. ANIMALS 71 dogs. PROCEDURES Urine samples (n = 84) were divided into 3 aliquots. One aliquot (reference culture) was plated on culture medium ≤ 1 hour after collection for quantitative culture and testing by standard laboratory methods, another was stored at 4°C for 24 hours (to mimic storage practices at primary care facilities) and then processed by standard methods, and the third was applied to a CCSP ≤ 1 hour after collection to be processed and interpreted according to manufacturer instructions. Results were compared with those for reference culture, which was used as the criterion reference standard. Sensitivity, specificity, positive and negative predictive values, and agreement between methods was evaluated. RESULTS 43 isolates (25 single and 9 multiple isolates) were identified in 34 reference cultures. All results for stored cultures were identical to those for reference cultures. Overall sensitivity of the CCSP method to detect bacteriuria was 93%, and specificity was 100%. Thirty-three of 43 (77%) and 19 of 33 (58%) CCSP bacterial isolates were correctly identified to the genus and species level, respectively. The CCSP antimicrobial susceptibility results matched those for reference cultures for 13 of 33 (39%) isolates evaluated. CONCLUSIONS AND CLINICAL RELEVANCE Limitations of the CCSP method included inaccuracy of some antimicrobial susceptibility test results and failure to correctly identify bacteriuria in some dogs. ( J Am Vet Med Assoc 2017;251:922–928)
B
acterial UTIs are common in dogs. Authors of 1 report1 estimated that 14% of dogs have ≥ 1 UTI during their lifetime. A UTI is diagnosed on the basis of clinical signs consistent with the condition and documented bacteriuria. The gold standard for identification of bacteriuria is a positive result for quantitative bacterial culture of urine (≥ 103 CFUs of bacteria/mL) with concurrent antimicrobial susceptibility testing to optimize treatment and minimize development of antimicrobial resistance.2 The immediate plating of urine after collection for bacterial culture is recommended to minimize false-negative or false-positive results.3,4 The practical alternative to immediate culture has been urine storage, either at 4°C without the adABBREVIATIONS CCSP Compartmentalized culture and antimicrobial susceptibility testing plate NPV Negative predictive value PPV Positive predictive value UTI Urinary tract infection 922
dition of a preservative or at room temperature with the addition of boric acid, followed by plating of the sample for culture ≤ 24 hours after collection.5–8 Although urine is a good storage medium, both strategies have limitations, especially for the preservation of Escherichia coli, one of the most common urinary pathogens isolated.3,8 Urine storage may alter bacterial growth, especially if a sample is not refrigerated within 1 to 2 hours after collection.3,8 The use of boric acid as a preservative has been reported to have an inhibitory effect on bacterial growth in urine samples from people7,9 and dogs.10 Unless a veterinary practice has a microbiology laboratory, most veterinarians must store urine samples and then ship them ≤ 24 hours after collection to a commercial laboratory for bacterial culture and antimicrobial susceptibility testing. Anecdotal information from primary care veterinary practices indicates bacterial urine cultures are often not pursued because of the associated cost and the perceived prolonged turnaround time. In addition, the
JAVMA • Vol 251 • No. 8 • October 15, 2017
Small Animals
impression of some practitioners is that there can be discordance between the clinical history or clinical signs of a dog suspected to have a UTI and the results obtained from urine cultured after storage for 24 hours (ie, a presumed false-negative result associated with urine storage). A quantitative in-clinic urine culture that could be performed immediately after sample collection would be an ideal method, but for most practice situations, this has not been practical or economical. Historically, point-of-care urine culture kits have been marketed and used in human medicine for the assessment of bacteriuria.11–13 Recently, similar urine culture kits (eg, a dipstick paddle system14 and CCSPs15,16) have been described for confirmation of bacteriuria and facilitation of bacterial identification in dogs. The CCSPs also provide antimicrobial susceptibility results as a means to aid antimicrobial selection. These systems were designed in part to address the limitations of having to send out urine samples for microbial culture and antimicrobial susceptibility testing by a microbiology laboratory. To the authors’ knowledge, independent field investigations on the clinical performance of these veterinary point-of-care urine culture systems have been limited to 2 studies14,16 published prior to this investigation and 1 study15 that was published after the completion of the study reported here. Those studies compared canine urine culture results obtained through a university microbiology laboratory by standard quantitative methods with results of either the dipstick paddle14 or CCSP15,16 systems. The reported sensitivities and specificities for the CCSP system were 81% to 83% and 99% to 100%, respectively, and those for the dipstick paddle system were 92% and 99%, respectively. The purpose of the study reported here was to evaluate use of the CCSP system to identify bacteriuria and the uropathogens present in a population of dogs at risk for bacteriuria by comparing test results with those obtained by standard microbiological methods. An additional objective was to determine whether antimicrobial susceptibility data obtained with the CCSP system would result in appropriate clinical antimicrobial selection for dogs from which a single uropathogen was isolated. We also sought to determine whether urine storage with refrigeration for 24 hours (to mimic the typical situation in primary care practices) would affect the results of culture by standard microbiological methods.
Materials and Methods Dogs and samples
Urine samples from client-owned dogs evaluated at UW Veterinary Care from June 11, 2013, to December 30, 2013, were used in this prospective, observational study. Only samples from dogs considered to be at risk for bacteriuria on the basis of history
or clinical findings and for which the clinician pursued a microbial culture and antimicrobial susceptibility testing as part of the clinical work-up were included. Dogs of any age, breed, or sex that had urine collected by cystocentesis, catheterization, or clean midstream catch of a voided sample were eligible for inclusion. Dogs were excluded from the study if the sample volume was insufficient (ie, < 5 mL) to ensure adequate amounts for both the clinically requested diagnostics and the tests performed for research purposes. All study protocols were reviewed, approved, and conducted in accordance with the requirements of the University of Wisconsin Animal Care and Use Committee. Data collected and recorded for all dogs enrolled in the study included age, sex, breed, weight, medical history, medications being administered at time of enrollment and antimicrobial treatments given in the previous 7 days, clinical diagnosis, and treatments recommended by the clinician responsible for the case (including antimicrobial administration, when prescribed).
Sample handling
Each urine sample was divided into 3 aliquots and placed into sterile plastic containers without additives after collection. Less than 1 hour after sample collection, the first aliquot was plated on standard urine culture medium for quantitative culture and antimicrobial susceptibility testing by standard laboratory methods for diagnostic purposes as requested by the attending clinician. These methods (with samples plated ≤ 1 hour after collection) were considered the clinical criterion reference standard methods for study purposes. Immediately after plating of this sample (ie, the reference culture sample), the remaining aliquots were processed. One of the aliquots was refrigerated at 4°C for subsequent culture and testing by standard laboratory methods, and the other was applied to a CCSP.a
Reference cultures
Quantitative bacterial culture consisted of inoculating 10 µL of urine collected by clean midstream catch or 100 µL of urine collected by cystocentesis onto both trypticase soy agar supplemented with sheep blood (to a 5% concentration) and eosin–methylene blue agar.b Urine cultures were evaluated for growth after 24 and 48 hours of incubation at 36°C in 5% CO2 atmosphere. Growth was quantitated as the number of CFUs per mL. Bacterial growth cutoffs were defined for each urine collection method used on the basis of recommendations in human and veterinary medical literature.17–21 For the present study, a positive result from a voided sample was defined as 1 to 3 microbial species present at concentrations > 10,000 CFUs/mL. The presence of > 3 species of microorganisms or concentrations < 10,000 CFUs/mL indicated probable urethral contamination, and results of these cultures
JAVMA • Vol 251 • No. 8 • October 15, 2017
923
Small Animals
were considered to represent unremarkable growth. Any growth of microorganisms from urine collected by cystocentesis was considered clinically relevant unless it was suggestive of gastrointestinal content contamination during cystocentesis. All cultures that yielded unremarkable growth were scored as having negative results for purposes of the study. Identification of bacterial isolates was made on the basis of colony type and morphology, Gram-staining characteristics, and standard biochemical test results, and information obtained with a commercial identification system.c Minimal inhibitory concentrations were determined for these isolates by use of a commercial systemc,d according to the manufacturer’s recommendations. Quality control was performed and antimicrobial susceptibility results were interpreted in accordance with Clinical Laboratory Standards Institute guidelines.22–24 Intermediate susceptibility interpretations were considered resistant for purposes of comparison in this study.
Storage of samples prior to culture
The designated urine sample aliquot was refrigerated at 4°C for 24 hours to mimic the standard procedure for storing urine before and during shipment to a microbiology laboratory. After the 24-hour storage time, the urine was cultured by the same standard microbiological methods as described for the reference culture samples. Bacterial quantitation, isolate identification, and antimicrobial susceptibility testing for these samples (deemed stored sample cultures) were performed as described for reference cultures by personnel in the microbiology section of the laboratory.
CCSP testing
The CCSPa used was designed to aid in the diagnosis of UTIs in dogs by providing semiquantitative bacterial counts and antimicrobial susceptibility results for uropathogens isolated from fresh samples or samples stored at 2° to 8°C for < 48 hours.25 All samples used with the CCSPs in the present study were fresh (applied ≤ 1 hour after collection). The plates were used in accordance with the manufacturer’s package insert25 prior to their expiration date. Plates were stored at 2° to 8°C and warmed to room temperature prior to use. Culture plating was performed by 1 investigator (AU) over the course of the study. Each plate consisted of a Petri dish containing ready-to-use Mueller-Hinton II agar divided into 1 large compartment free of antimicrobials (used for colony counts) and 5 smaller compartments, each containing 1 antimicrobial (ampicillin, amoxicillin-clavulanate, cephalothin, enrofloxacin, or trimethoprim-sulfamethoxazole; used for susceptibility testing of microbes).25 Urine samples were plated on CCSPs (1 plate/ sample) according to the manufacturer’s recommendations. Briefly, urine (approx 1 to 2 mL) was flooded onto the plate for 3 to 5 seconds in a manner that ensured all compartments were coated, and any ex924
cess urine was poured off the plate. The plate was then incubated in ambient atmospheric conditions at 35°C and checked for growth by gross visual examination after 24 hours of incubation by the same investigator (AU), who was unaware of the reference culture results. The plate was incubated for an additional 24 hours if no growth was observed at the first inspection. Clinically relevant bacterial growth was determined by visual inspection of the large counting compartment, and the results for each sample were determined according to the interpretive guidelines provided by the manufacturer. Briefly, bacterial isolates were identified by color changes of the medium specific to bacterial species, and an estimate of the number of CFUs was made on the basis of colony density as described and illustrated by the manufacturer’s instructions.25 Total colony counts from voided samples and urine collected by cystocentesis were interpreted in a manner similar to that used for reference culture results. For plates deemed to have clinically relevant bacterial growth, the antimicrobial susceptibility compartments of the plate were evaluated and results were recorded. Growth in each of these compartments was compared with that in the larger antimicrobial-free compartment to determine whether the isolate was susceptible or resistant to the antimicrobials included in the plate (an interpretation of intermediate could not be determined by this method). An isolate was considered susceptible if there was no growth or resistant if there was growth in a particular antimicrobial compartment.
Statistical analysis
Results of a power calculation performed on the basis of data from a study11 investigating use of a similar CCSP-based urine culture system in human patients indicated that 70 urine samples would be needed to achieve 80% power to detect bacteriuria in dogs. The Cohen κ coefficient was calculated (with 95% confidence interval) as a quantitative measure of the magnitude of agreement between culture methods beyond that expected by chance. Positive and negative κ coefficients indicate agreement greater than or less than that expected by chance alone, respectively. The degree of agreement was interpreted as described elsewhere (1 = perfect; 0.80 to 0.99 = very good; 0.6 to 0.79 = good; 0.4 to 0.59 = moderate; 0.2 to 0.39 = fair; and 0 to 0.2 = poor).26 The sensitivity, specificity, PPV, NPV, and accuracy of the CCSP method (with 95% confidence interval) were calculated with standard equations.26 For these indices of clinical performance, 1 sample/dog was included in the analyses. In cases where multiple samples had been obtained from a given dog and cultures were repeated, only the results of the first culture were included in the statistical analysis.
Results Eighty-four urine samples collected by cystocentesis (n = 48), catheterization (4), or clean mid-
JAVMA • Vol 251 • No. 8 • October 15, 2017
Small Animals
stream catch of a voided sample (32) were included in the study. Samples were collected from 71 dogs (25 males [22 neutered and 3 sexually intact] and 46 females [40 spayed and 6 sexually intact]). There were 63 purebreds and 8 mixed-breed dogs; the 5 breeds most frequently represented were Labrador Retriever (n = 7), Beagle (5), Bichon Frise (5), American Pit Bull Terrier (4), and Dachshund (4). The median age of dogs was 8.6 years (range, 1 to 15 years), and median weight was 19 kg (42 lb; range, 4.6 to 96.9 kg [10.1 to 213.2 lb]). Clinical diagnoses included UTI (n = 35), neoplasia (9 [3 and 6 involving and not involving the urinary bladder, respectively]), proteinuria (6), diabetes mellitus (3), acute kidney injury (3), and urinary incontinence (3). Twelve dogs had other clinical diagnoses, including inappropriate urination attributed to behavioral causes, hematuria of undetermined cause, vaginitis, renal dysplasia, chronic kidney disease, hyperadrenocorticism, hypoadrenocorticism, protein-losing nephropathy, nephrolithiasis, pyogranulomatous lymphadenitis, sepsis, and uveitis. At the time of urine culture, 21 dogs were receiving or had received antimicrobial treatment within the past 7 days. These treatments included β-lactams (n = 13), fluoroquinolones (4), trimethoprim-sulfamethoxazole (2), clindamycin (1), and metronidazole (1).
Bacterial isolates
Culture results attained by standard microbiological methods (for samples plated ≤ 1 hour after collection [reference cultures] and those plated after refrigeration for 24 hours [stored sample cultures]) and by use of the CCSP were summarized (Table 1). The results for stored sample cultures were identical to those of the reference cultures for all tests, including detection of a positive result, genus and species
identification of isolates, and results of antimicrobial susceptibility testing. Compared with the 43 isolates identified in reference cultures, the CCSP correctly identified 33 (77%) isolates to the genus level, with 19 of 33 (58%) isolates identified to the species level and 14 of 33 (42%) identified only to the genus level (Table 1). The 33 isolates detected by CCSP included 13 of 19 (68%) gram-positive and 20 of 24 (83%) gram-negative isolates, compared with reference culture results. Isolates identified in reference cultures included E coli (n = 18), Staphylococcus pseudintermedius (7) and methicillin-resistant S pseudintermedius (2), Streptococcus canis (7), Proteus mirabilis (3), Enterobacter cloacae (2), Enterococcus faecium (2), Enterococcus faecalis (1), and Pseudomonas aeruginosa (1). Results for the same samples cultured by use of CCSPs revealed identification of E coli (n = 16), E cloacae (1), E faecalis (1), and P aeruginosa (1). Isolates identified only at the genus level, as allowed for by the CCSP method, included Staphylococcus spp (7), Streptococcus spp (3), Proteus spp (2), and Enterococcus spp (2). Cohen κ analysis confirmed perfect agreement between results (positive vs negative) obtained with reference cultures and stored sample cultures (κ = 1). Agreement between results (positive vs negative) for CCSPs and reference cultures was very good when all dogs (κ = 0.94) and only dogs that had a single bacterial isolate identified in reference cultures (κ = 0.93) were included in the analysis. Agreement for CCSPs and reference cultures was fair (κ = 0.36) for samples from dogs in which multiple bacterial isolates were identified in reference cultures (Table 2). The sensitivity, specificity, PPV, NPV, and accuracy of the CCSP method to detect clinically relevant
Table 1—Summary of results for 84 urine samples cultured and tested by standard laboratory methods (with samples plated ≤ 1 hour after collection [clinical reference culture] or after storage at 4°C for 24 hours [stored sample culture]) and by use of a CCSP system. Variable
Clinical Stored reference culture sample culture
Culture result (No. of samples) Positive 34 34 Single isolate 25 25 Multiple isolates 9 9 Negative 50 50 Bacterial isolate identification No. of isolates 43 43 Gram-positive 19 19 Gram-negative 24 24 Genus and species identified 43 43 Genus identified only 0 0
CCSP culture 30 27 3 54 33 13 20 19 14
Urine was collected from 71 dogs considered to be at risk for bacteriuria. Samples were collected by cystocentesis (n = 48), catheterization (4), or clean midstream catch of a voided sample (32) and divided into 3 aliquots (1/method). A positive culture result was defined as clinically relevant growth; for the reference culture method, cutoffs were determined as described elsewhere,17–21 and for the CCSP, this was determined according to the manufacturer’s instructions for use of the plates. The reference culture result was used for diagnostic purposes for dogs in the study.
JAVMA • Vol 251 • No. 8 • October 15, 2017
925
Small Animals
Table 2—Sensitivity, specificity, PPV, NPV, and accuracy for the use of CCSPs to identify bacteriuria in urine samples from 71 dogs, compared with results for reference cultures and assessment of agreement between the 2 methods as measured by Cohen k analysis. PPV Test
NPV
No. of dogs Sensitivity (%) Specificity (%) Accuracy (%) TP/TP+FP % TN/TN+FN %
Bacteriuria (all dogs) Bacterial isolate identified (genus) from cultures with a single isolate† Bacterial isolate identified (genus) from cultures with multiple isolates‡
κ*
71 63
93 (77–99) 90 (69–98)
100 (92–100) 100 (93–100)
97 (91–100) 97 (91–100)
27/27 19/19
100 (87–100) 100 (82–100)
42/44 42/44
95 (85–99) 0.94 (0.86 to 1) 96 (87–99) 0.93 (0.84 to 1)
50
25 (3.2–65)
100 (92–100)
88 (79–97)
2/2
100 (16–100)
42/47
88 (75–95) 0.36 (–0.01 to 0.73)
Data are shown as value (95% confidence interval) where appropriate. When multiple samples were cultured for a given dog, only the first sample was included in these analyses. *The κ coefficient was used to determine the degree of intermethod agreement as follows: 1= perfect, 0.80 to 0.99 = very good, 0.6 to 0.79 = good, 0.4 to 0.59 = moderate, 0.2 to 0.39 = fair, and 0 to 0.2 = poor. †Eight samples that had multiple isolates identified by the reference culture method were excluded from this analysis. ‡For dogs with multiple isolates, calculations were based on correct versus incorrect identification of all isolates. FN = Number of false-negative results. FP = Number of false-positive results. TN = Number of true-negative results. TP = Number of true-positive results. See Table 1 for remainder of key.
Table 3—Empirical antimicrobial treatments and antimicrobial susceptibility profiles for isolates from urine samples of 9 dogs that had positive culture results for a single isolate and discrepancies in susceptibility data between the reference culture and CCSP methods. Antimicrobial susceptibility Empirical Isolate and treatment Dog No. culture method AMP AMC CEP ENR TMS prescribed 1 2 3* 4* 5* 6 7 8* 9*
E faecalis RC CCSP E coli RC CCSP E coli RC CCSP E coli RC CCSP E coli RC CCSP S pseudintermedius RC CCSP E coli RC CCSP E coli RC CCSP E coli RC CCSP
S S — I — AMX S S R R S S S S S S AMC S S R S S S S I S S AMX R S S S S S S S S S AMX R R R S S S S I S S AMC R R R S S R S S S S AMC S S S S S R S I S S CIP R R R S S S S I S S CEF R R R S S R S S S R AMC R R R S S
Antimicrobial susceptibility testing for reference cultures was performed by standard microbiology methods with results (susceptible [S], intermediate [I], or resistant [R]) interpreted according to Clinical Laboratory Standards Institute guidelines.21–23 Results for testing with CCSPs were interpreted according to the manufacturer’s instructions. *Reliance on CCSP results alone for these dogs could have resulted in an unnecessary change in empirical antimicrobial treatment. — = Not applicable. AMC = Amoxicillin-clavulanate. AMP = Ampicillin. AMX = Amoxicillin. CEF = Cefpodoxime. CEP = Cephalothin. CIP = Ciprofloxacin. ENR = Enrofloxacin. RC = Reference culture. TMS = Trimethoprim-sulfamethoxazole.
bacterial growth in samples from all dogs, dogs with single bacterial isolate identified, and dogs with multiple bacterial isolates identified were summarized (Table 2). One or more isolates were missed by use 926
of the CCSP in 6 of 9 samples from dogs confirmed to have mixed bacterial infections and 4 of 25 samples from dogs with a single isolate identified (including multiple cultures from individual dogs).
JAVMA • Vol 251 • No. 8 • October 15, 2017
Small Animals
Antimicrobial susceptibility
Comparison of antimicrobial susceptibility testing results by standard microbiological methods (for bacterial isolates from reference cultures) versus those obtained with CCSPs yielded the same results for 13 of 33 (39%) isolates that were identified by both methods. For the remaining 20 (61%) isolates, the in vitro antimicrobial susceptibility results by CCSP conflicted with those for the reference method. These disparate results were attributed to identification of different antimicrobial susceptibility interpretations (susceptible vs resistant) for a particular antimicrobial with the 2 methods (n = 18) or inappropriate antimicrobial testing with CCSPs (2; both Enterococcus spp) owing to fixed antimicrobial testing. Susceptibility testing results for E coli isolated from 16 samples by use of CCSPs were incorrect for ≥ 1 of the β-lactam antimicrobials tested (ampicillin [n = 8], amoxicillin-clavulanic acid [8], or cephalothin [4]). Four of 7 Staphylococcus spp isolates were incorrectly interpreted as being susceptible to ampicillin by CCSP, and 1 of 3 isolates of Streptococcus spp was incorrectly identified as resistant to enrofloxacin. When antimicrobial susceptibility test results for samples that had only 1 isolate identified by both methods (n = 21) were compared, findings were the same for 11 (52%) and conflicting for 10 (48%). Disparate results for 8 of 10 isolates were primarily attributed to incorrect interpretation results for susceptibility to β-lactams. Review of these cases revealed that empirical antimicrobial treatment was initiated for 9 dogs with clinically relevant bacteriuria caused by single isolates that also had discrepancies in antimicrobial susceptibility interpretations between the 2 methods. On the basis of results for the CCSP testing alone, empirical treatment could have been unnecessarily changed for 5 of 9 dogs. In comparison, empirical treatment might have been changed for only 1 of these 9 dogs on the basis of reference culture results (Table 3).
Discussion Results of the present study confirmed that storage of canine urine samples at 4°C for 24 hours did not affect the results of culture and susceptibility testing of isolates by standard microbiological methods. Results of bacterial genus and species identification as well as antimicrobial susceptibility testing for isolates from stored sample cultures were identical to those for reference cultures (plated ≤ 1 hour after sample collection for culture and testing by the same methods), supporting that storage of urine overnight at 4°C is a viable clinical alternative when immediate urine culture is not possible. In addition, samples cultured and evaluated on CCSPs had very good agreement with reference cultures for identification of clinically relevant bacterial growth or for single-isolate infections, but only fair agreement for polymicrobial infections, in which isolates were sometimes not identified. On the basis of the calculated overall sensitivity (93%) and specificity (100%) of the CCSP method in this study, growth on the CCSP culture is a good indicator of bacteriuria (PPV, 100%); however, a no-growth CCSP re
sult could miss identifying bacteriuria (NPV, 95%) in this population of dogs at risk for the condition. For dogs with positive bacterial culture results, our findings indicated that the antimicrobial susceptibility data from the CCSP culture should be interpreted cautiously when making therapeutic decisions, especially for the use of β-lactam antimicrobials. In this study, reliance on only the CCSP antimicrobial susceptibility results could have resulted in an unnecessary change or escalation (eg, use of fluoroquinolones) in antimicrobial administration for 5 of 9 dogs. Furthermore, of the 7 Staphylococcus spp isolates identified by both methods, the CCSP falsely indicated susceptibility to ampicillin for 4 isolates and missed detection of methicillin resistance for 2 isolates owing to a lack of oxacillin susceptibility testing by the CCSP plate. In addition, reliance on the CCSP antimicrobial susceptibility results for E faecalis isolated from 1 sample could have led to the inappropriate selection of trimethoprim-sulfamethoxazole for treatment; although an isolate may appear susceptible to the drug in vitro, it is not clinically effective against enterococci.22 These findings suggested that use of antimicrobial susceptibility data obtained by interpretation of CCSP results may not promote improved antimicrobial use and could potentially contribute to treatment failure or increased antimicrobial resistance of uropathogens. The primary advantage of CCSP use as a point-ofcare diagnostic aid is its minimal technical training and equipment requirements, the short set-up time needed after urine collection, and avoidance of the need for storage of urine before it can be cultured. The use of this inclinic method has the potential to be less expensive than shipping urine specimens to a microbiology laboratory for evaluation. Furthermore, plates can be evaluated 24 hours after culture is initiated and might provide some preliminary information to aid in case management, although results must be interpreted in the context of the patient’s clinical history and response to treatment. Some disadvantages of CCSP use are that plating requires a fairly large urine volume (minimum of 1.5 mL); antimicrobial susceptibility testing is limited to 5 antimicrobials, which may be inappropriate for some isolates; and susceptibility test results, as identified in this study and others,15,16 frequently conflict with those obtained by standard microbiological methods. In addition, the interpretation of the results relies on colorimetric changes, dependent on the evaluator’s subjective judgement for isolate identification. In the present study, accurate identification of isolates was subjectively difficult or impossible in some samples, resulting in inconsistencies between results for this method and that used for reference cultures, particularly in some samples with multiple bacterial species present, low bacterial counts, or a color change that was difficult to judge. The results of this study were similar to findings of another clinical study16 that evaluated use of CCSPs for assessment of canine urine samples. In that study, results obtained with the CCSP method were compared with results of quantitative urine culture performed by a university microbiology laboratory in samples plated
JAVMA • Vol 251 • No. 8 • October 15, 2017
927
Small Animals
≤ 24 hours after collection. The reported sensitivity and specificity of the test (evaluated in samples from 96 dogs) was 81% and 99%, respectively. The results of both (previous and present) studies supported that quantitative urine culture with antimicrobial susceptibility testing through a microbiology reference laboratory is a more accurate means to aid in the diagnosis of canine bacteriuria and direct antimicrobial selection. Limitations of the present study included the relatively small number of canine urine specimens of adequate volume for both clinical diagnostic testing and use for research purposes, which limited power of the study. The overall low number of samples that had positive culture results limited comparisons among methods for various bacterial species. Additionally, the underlying conditions identified in this group of dogs were diverse, and dogs with bacteriuria included patients with conditions other than uncomplicated UTIs. However, the diversity of patients included in the study reflects to some degree the clinical population of dogs for which urine culture may be pursued. Finally, the handling of samples that were stored with refrigeration prior to culture might not have adequately represented the conditions of urine storage and transportation in practice. The results of this study of a clinical population of dogs suspected to have bacteriuria did not support the use of the CCSP method as a replacement for standard microbiological methods for bacterial culture and antimicrobial susceptibility testing of isolates from canine urine samples. However, for clinical situations in which a quantitative culture is not available, knowledge of the clinical limitations of the CCSP method is important for clinicians considering a point-of-care bacterial culture method.
5.
6. 7. 8. 9. 10. 11. 12.
13. 14.
15.
16. 17.
Acknowledgments No third-party funding or support was received in connection with this study or the writing or publication of the manuscript. The authors declare that there were no conflicts of interest and issue no disclaimers. Presented in part as a poster at the American College of Veterinary Internal Medicine Forum, Nashville, Tenn, June 2014.
18.
Footnotes
20.
a. b. c. d.
Flexicult Vet, Atlantic Diagnostics, Miami, Fla. Remel, Lenexa, Kan. Vitek 2, BioMérieux, Marcy l’Étoile, France. TREK Diagnostic Systems, Cleveland, Ohio.
References
21. 22.
23.
1. Ling GV. Therapeutic strategies involving antimicrobial treatment of the canine urinary tract. J Am Vet Med Assoc 1984;185:1162–1164. 2. Weese JS, Blondeau JM, Boothe D, et al. Antimicrobial use guidelines for treatment of urinary tract disease in dogs and cats: antimicrobial guidelines working group of the international society for companion animal infectious diseases. Vet Med Int 2011;2011:263768. 3. Hindman R, Tronic B, Bartlett R. Effect of delay on culture of urine. J Clin Microbiol 1976;4:102–103. 4. Perrin J, Nicolet J. Influence of the transport on the outcome of the bacteriological analysis of dog urine compari-
928
19.
24.
25. 26.
son of three transport tubes. Zentralbl Veterinarmed B 1992;39:662–667. Horton JA III, Kirshblum SC, Linsenmeyer TA, et al. Does refrigeration of urine alter culture results in hospitalized patients with neurogenic bladders? J Spinal Cord Med 1998;21:342–347. Lum KT, Meers PD. Boric acid converts urine into an effective bacteriostatic transport medium. J Infect 1989;18:51–58. Meers PD, Chow CK. Bacteriostatic and bactericidal actions of boric acid against bacteria and fungi commonly found in urine. J Clin Pathol 1990;43:484–487. Padilla J, Osborne CA, Ward GE. Effects of storage time and temperature on quantitative culture of canine urine. J Am Vet Med Assoc 1981;178:1077–1081. Gillespie T, Fewster J, Masterton RG. The effect of specimen processing delay on borate urine preservation. J Clin Pathol 1999;52:95–98. Rowlands M, Blackwood L, Mas A, et al. The effect of boric acid on bacterial culture of canine and feline urine. J Small Anim Pract 2011;52:510–514. Blom M, Sorensen TL, Espersen F, et al. Validation of Flexicult SSI-urinary kit for use in the primary health care setting. Scand J Infect Dis 2002;34:430–435. Scarparo C, Piccoli P, Ricordi P, et al. Evaluation of the DipStreak, a new device with an original streaking mechanism for detection, counting, and presumptive identification of urinary tract pathogens. J Clin Microbiol 2002;40:2169–2175. Winkens R, Nelissen-Arets H, Stobberingh E. Validity of the urine dipslide under daily practice conditions. Fam Pract 2003;20:410–412. Ybarra WL, Sykes JE, Wang Y, et al. Performance of a veterinary urine dipstick paddle system for diagnosis and identification of urinary tract infections in dogs and cats. J Am Vet Med Assoc 2014;244:814–819. Guardabassi L, Hedberg S, Jessen LR, et al. Optimization and evaluation of Flexicult(R) Vet for detection, identification and antimicrobial susceptibility testing of bacterial uropathogens in small animal veterinary practice. Acta Vet Scand 2015;57:72. Olin SJ, Bartges JW, Jones RD, et al. Diagnostic accuracy of a point-of-care urine bacteriologic culture test in dogs. J Am Vet Med Assoc 2013;243:1719–1725. Burd EM, Kehl KS. A critical appraisal of the role of the clinical microbiology laboratory in the diagnosis of urinary tract infections. J Clin Microbiol 2011;49:S34–S38. Carter JM, Klausner JS, Osborne CA, et al. Comparison of collection techniques for quantitative urine culture in dogs. J Am Vet Med Assoc 1978;173:296–298. Comer KM, Ling GV. Results of urinalysis and bacterial culture of canine urine obtained by antepubic cystocentesis, catheterization, and the midstream voided methods. J Am Vet Med Assoc 1981;179:891–895. Sanderson SL. Urinary system introduction. In: Kahn C, ed. The Merck veterinary manual. 10th ed. Whitehouse Station, NJ: Merck & Co, 2010;1380. Stamm WE, Wagner KF, Amsel R, et al. Causes of the acute urethral syndrome in women. N Engl J Med 1980;303:409–415. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing. CLSI document M100–23. Wayne, Pa: Clinical and Laboratory Standards Institute, 2013. Clinical and Laboratory Standards Institute. Performance Standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals; approved standard. 4th ed. CLSI document VET01–A4. Wayne, Pa: Clinical and Laboratory Standards Institute, 2013. Clinical and Laboratory Standards Institute. Performance Standards for antimicrobial disk and dilution susceptibility tests for bacteria isolated from animals. CLSI document VET01–S2. Wayne, Pa: Clinical and Laboratory Standards Institute, 2013. Flexicult vet [package insert]. Miami, Fla: Atlantic Diagnostics, 2013. Altman D. Practical statistics for medical research. London: Chapman & Hall, 1997.
JAVMA • Vol 251 • No. 8 • October 15, 2017
