Review of Enterococci Isolated from Canine and Feline Urine Specimens from 2006 to 2011

Introduction

Enterococci are gram-positive commensal bacteria in the gastrointestinal tract of mammals. Enterococci can also act as opportunistic pathogens causing substantial systemic disease, and this genus is now labeled as a leading cause of hospital-acquired infections in human healthcare.¹ Despite the clinical prevalence and vast amount of published information regarding enterococci in human healthcare, the prevalence, clinical significance, and optimal management of enterococcal infections in veterinary patients have not been fully investigated.

In veterinary medicine, urine cultures with enterococcal growth are seen routinely, representing 8.5–24% of positive canine urine cultures and 5–27% of positive feline urine cultures.^2–7 Management of enterococcal urinary tract infections (UTIs) can be challenging and is frequently complicated by the presence of mixed and complicated infections and multidrug resistance.^7–9 Optimal therapeutic strategies for veterinary patients with positive enterococcal cultures remain undefined, including what patients require therapy and what antibiotics are most likely to be effective. Circumstances when enterococcal growth from a canine or feline urine specimen should be classified as asymptomatic colonization are currently undefined in the veterinary literature. Determining if factors such as enterococcal species isolated and sediment changes (e.g., hematuria, pyuria) are associated with presence of clinical signs of a UTI may help define asymptomatic colonization with enterococci versus true UTI and to clarify veterinarians' understanding and management of enterococcal UTIs.

The primary goals of this study were to determine the proportion of aerobic urine cultures from dogs and cats analyzed between 2006 and 2011 at the Kansas State University Veterinary Health Center that were positive for growth of enterococci and of those that yielded significant bacteriuria to determine if associations existed between the presence of clinical signs or urine sediment findings (i.e., bacteriuria, pyuria) and various factors, including species of enterococci isolated, number of bacterial species isolated (Enterococcus only versus mixed Enterococcus infections) and simple versus complicated infections. Another goal was to determine if an association existed between multidrug resistant isolates and enterococcal species. It was hypothesized that the presence of clinical signs consistent with lower urinary tract disease would be associated with simple UTIs and that multidrug resistant isolates would occur more commonly with Enterococcus (E.) faecium than E. faecalis.

Materials and Methods

This was a retrospective review of aerobic urine cultures and corresponding urinalyses submitted to the Kansas State Veterinary Diagnostic Laboratory from canine and feline patients of the Kansas State University Veterinary Health Center between 2006 and 2011. Due to the retrospective nature of the study, urine specimen handling prior to delivery at the laboratory was not standardized; however, hospital policy included submission of urine samples immediately during daytime hours and, based on clinician preference, either having a student laboratory technician plate the sample after hours (7 days/wk) or refrigeration (<15 hr, 7 days/wk) in a red top tube (i.e., no enrichment or medium) overnight when the laboratory was closed with submission as early as possible when the laboratory reopened.

Urine specimens submitted during that time period were initially plated on trypticase soy agar with 5% sheep blood and MacConkey agar using a 10 μL calibrated loop. Enrichment was also performed routinely at the laboratory; however, enriched cultures were not included in the analyzed dataset for this study due to the definition of significant bacteriuria. Cultures were incubated overnight (18–24 hr) at 37°C in 5% CO₂. Suspect colonies were struck for isolation onto nonselective agar. Prior to June 2010, isolates were identified using a combination of gram staining, catalase, and traditional biochemical testing.¹⁰ After June 2010, isolates were identified using a computerized identification system^a. Antimicrobial susceptibility was performed using microwell dilution testing^b throughout the survey period, according to published guidelines.¹¹ During the study period, there were no revisions in the Clinical and Laboratory Standards Institute's (CLSI's) recommendations for testing methodology or interpretations for enterococci.

Information collected from culture reports included presence and quantification of enterococcal growth, enterococcal species, susceptibility to antimicrobial agents, and coinfections with other bacterial species. Susceptibilities to aminoglycosides, cephalosporins, clindamycin, and trimethoprim-sulfa were not reported in this study as per the guidelines of the CLSI.¹² Multidrug resistance (MDR) was defined as resistance to three or more categories of antimicrobial agents [i.e., β-lactams (excluding cephalosporins); fluoroquinolones; erythromycin; chloramphenicol; tetracyclines].

Urinalyses performed at the Kansas State Veterinary Diagnostic Laboratory within 24 hr of culture submission were reviewed to determine whether bacteria (cocci) or WBCs were identified in urine sediment by trained laboratory technicians. For the purpose of this study, cocci were recorded as either present or not present and pyuria was defined as >5 WBCs/high-power field.⁶ Medical records were analyzed to determine presence of clinical signs (e.g., pollakiuria, hematuria, stranguria) that would support a UTI and for the presence of concurrent conditions or therapy that would suggest a complicated UTI. A simple, uncomplicated UTI was defined as a sporadic bacterial infection of the urinary bladder in an otherwise healthy individual with normal urinary tract anatomy and function. A complicated UTI was defined as either the presence of relevant comorbidities (e.g., diabetes mellitus, urinary or reproductive tract conformational abnormalities) or ≥3 UTI/yr.⁹

Results

Table 1 displays the breakdown of submitted urine cultures, urine cultures positive for any species of bacterial growth, urine cultures positive for enterococcal growth (including those with enrichment only), and urine cultures with significant enterococcal growth from urine collected by either cystocentesis or sterile catheterization in specimens from dogs and cats. Over the study period, an average of 15% of urine specimens with bacterial growth from 109 out of 745 dogs (14.6%; 95% CI, 9.7–19.5) and 29 out of 195 cats (14.9%; 95% CI, 10–19.8%) had enterococcal growth. Roughly one-third of those specimens [31 out of 109 dogs (28.4%; 95% CI, 14.6–42.2%) and 11 out of 29 cats (37.9%; 95% CI, 12.9–62.9%)] had significant enterococcal bacteriuria to be defined as a urinary tract infection.

TABLE 1 Number of Specimens Submitted for Urine Culture, Urine Cultures Positive for Bacterial Growth, Urine Cultures Positive for Enterococcal Growth, and Urine Cultures with Significant Urine Enterococcal Growth Collected by Cystocentesis or Sterile Catheterization from Dogs and Cats Between 2006 and 2011

View Fullscreen

Of the 29 feline urine cultures with enterococcal growth, 3 were from the same patient but only the first culture from that patient was included for analysis. Method of urine collection was unknown for 2 cultures, which were therefore excluded from further analysis. Of the 25 remaining cultures, 8 had ≥100,000 CFUs/mL urine (6 collected by cystocentesis and 2 collected by catheter), 3 had between 10,000 and 100,000 CFUs/mL urine (all 3 collected by cystocentesis), no cultures had between 1000 and 10,000 CFUs/mL urine, and 14 cultures had either <1000 CFUs/mL urine or nonnumeric descriptions of enterococcal growth.

All 11 feline isolates included in further analysis were identified as E. faecalis, and 9 out of 11 infections were considered complicated UTIs. Concurrent conditions that contributed to complicated infections included: uroliths (one cat), cystostomy tube (one cat), urethral catheterization following urethral rupture (one cat), spinal dysfunction (one cat), chronic kidney disease (six cats), and hyperthyroidism (two cats). Three cats had more than one concurrent disorder.

Clinical signs consistent with lower urinary tract disease were observed in 5 out of 11 cats (45.5%). Eight specimens had E. faecalis as the only bacterium isolated and three cultures had mixed bacterial growth (E. faecalis with one each of Enterobacter cloacae, Pseudomonas aeruginosa, or Escherichia coli). Two-thirds of cats with mixed bacterial cultures had clinical signs consistent with a UTI, whereas 4 out of 8 cats with only E. faecalis had clinical signs. No relationship was found between the presence of clinical signs and a simple versus complicated infection (P = .454); however, an additional 30 cases would have been necessary to provide a power of 80% to detect such an association.

Of the 11 feline urine cultures analyzed, 8 had concurrent urinalyses available. Cocci were identified microscopically in all 8 urine sediments and 5 out of 8 had pyuria. No relationship was found between the presence of clinical signs and either presence or absence of pyuria (P = .464); however, an additional 27 cases would have been necessary to provide a power of 80% to detect such an association.

The percent of feline urinary isolates resistant to tested antimicrobial agents is shown in Table 2. Of the feline enterococcal urine isolates, antimicrobial resistance occurred most commonly with oxacillin [minimum inhibitory concentration (MIC) >4 μg/mL for all 11 isolates] and enrofloxacin (7 out of 11 isolates). For enrofloxacin testing of the remaining isolates, 2 had MICs ≤0.5 μg/mL and 2 had MICs of 1 μg/mL. Six feline isolates were classified as MDR, with all six resistant to oxacillin. Four out of 6 isolates were resistant to enrofloxacin, 4 out of 6 were resistant to erythromycin, 3 out of 6 were resistant to chloramphenicol, and 3 out of 6 were resistant to tetracycline.

TABLE 2 Number and Percent of Enterococcal Isolates from Feline and Canine Urine Specimens that were Resistant to Listed Antimicrobial Agents as Determined by Microwell Dilution Testing*

View Fullscreen

Of the feline specimens listed above, two cats (one 2 yr old and one 7 yr old spayed female) had urine specimens with >100,000 CFU/mL E. faecalis growth and did not have any identified concurrent disease or predisposing factors. Both specimens were collected by cystocentesis. One was mixed with Escherichia coli, and one had enterococci alone. Both had many cocci on sediment exam; one had pyuria and one only had occasional WBCs. Owners of those cats reported inappropriate urinating around the house and malodorous urine as clinical signs. The E. faecalis isolate from one of these cats (the one with only enterococci) was MDR.

Six canine patients had more than one urine sample culture positive for enterococcal growth, and only the first culture from each of these patients was considered for further analysis. Three canine urine specimens with >100,000 CFUs/mL were excluded because urine was collected by clean free catch. Four canine urine culture specimens had unreported collection technique and were excluded as well. Of the remaining samples, 31 aerobic cultures of canine urine were included for further analysis. Of those, 14 had ≥100,000 CFUs/mL urine (12 collected by cystocentesis, 2 collected by sterile catheterization), 15 had between 10,000 and 100,000 CFUs/mL urine (12 collected by cystocentesis, 3 collected by sterile catheterization), and 2 cultures had between 1000 and 10,000 CFUs/mL urine (collected by cystocentesis).

The 31 analyzed canine urine culture specimens were comprised of 24 E. faecalis, 4 E. faecium, 1 E. durans, and 2 uncharacterized Enterococcus spp. The majority (26 out of 31, 83.8%) were considered complicated infections, including 10 dogs with neurologic dysfunction; 4 dogs with diabetes mellitus; 4 dogs with sphincter incontinence; 4 dogs receiving corticosteroid therapy; 3 dogs having uroliths; 2 dogs with bladder neoplasia; 2 dogs with recessed vulva; 2 dogs with indwelling urinary catheters; 1 dog with chronic kidney disease; and 1 dog with hyperadrenocorticism. Several dogs had more than one complicating concurrent condition. Of the 31 culture specimens, 23 were enterococci-only specimens and 8 were specimens with enterococci mixed with another bacterial species. Clinical signs consistent with lower urinary tract disease were seen in 15 out of 31 dogs (48.4%) with enterococcal UTIs, including 3 out of 8 dogs with mixed bacterial specimens and 12 out of 23 dogs with specimens growing only enterococci. Of the mixed bacterial specimens, 4 out of 8 contained E. faecalis (1 mixed with Klebsiella pneumoniae and 3 mixed with Escherichia coli), 3 out of 8 contained E. faecium (1 mixed with Enterobacter cloacae and 2 mixed with Escherichia coli), and 1 out of 8 contained an uncharacterized Enterococcus spp. mixed with a Staphylococcus spp. No association was identified between either the presence or absence of clinical signs and enterococcal species (P = .394); however, an additional 134 cases would have been necessary to provide a power of 80% to detect such an association. Similarly, no association was identified between presence or absence of clinical signs and simple versus complicated infections (P = .171); however, an additional 69 cases would have been necessary to provide a power of 80% to detect such an association.

A concurrent urinalysis was available for 25 out of 31 canine urine culture reports. Of those urinalyses, cocci were noted microscopically in 21 out of 25 urine sediments, and pyuria was documented in 15 out of 25. Presence of pyuria in canine urine sediments was associated with growth of only Enterococcus (versus mixed bacterial species) on aerobic urine culture (P = .005). Of the 15 canine patients with pyuria identified on urine sediment exam, 100% had only an Enterococcus isolated from their urine. Identification of cocci on sediment exam was not significantly associated with only Enterococcus infection versus mixed bacterial species (P = .166); however, an additional 45 cases would have been necessary to provide a power of 80% to detect such an association.

Percent resistance of canine enterococcal isolates to tested antimicrobial agents is also displayed in Table 2. Of the canine E. faecalis isolates, antimicrobial resistance was most commonly seen to oxacillin (MIC > 4 μg/mL for all 24 isolates) and to enrofloxacin (10 out of 24 isolates). For enrofloxacin testing of the 14 remaining E. faecalis isolates, 1 had a MIC of 0.5 μg/mL and 13 had a MIC of 1 μg/mL. Seven canine isolates displayed MDR. E. faecium isolates (3 out of 4) were 4.5× more likely to be MDR than E. faecalis isolates (4 out of 24; P = .038; 95% CI, 1.56–12.97).

Discussion

In both canine and feline patients of the authors' veterinary teaching hospital, approximately 15% of urine specimens submitted for aerobic culture that were positive for bacterial growth identified enterococci, which is consistent with other prevalence studies and underscores the relevance of this genus in UTI discussions.^2–7 When considering quantitative urine culture in light of urine collection method to classify UTIs as significant, E. faecalis was the most commonly cultured enterococcal species from dogs and cats in this study, followed by E. faecium. In contrast, another study performed at Michigan State University found that enterococcal UTIs in dogs were most commonly caused by E. faecium (37%), followed by E. gallinarum (31%) and E. faecalis (20%).⁸ In that study, method of urine collection and quantification of enterococci (CFU/mL) were not provided, making it unclear if isolates included in that study would have fit the criteria for significant bacteriuria used in the present study. Other possible reasons for the discrepant percentage of species identified in the Michigan State study could include differences in geography, time period (the Michigan State isolates were from 1996 to 1998), methodology for species identification, or random error from low sample sizes.

The majority of feline urine specimens with significant enterococcal cultures originated from cats with concurrent disease; however, two cats under the age of 10 yr had enterococcal UTIs without evidence of either concurrent disease or predisposing factors. The ability to cause disease in otherwise healthy young or middle age cats supports enterococci being a true uropathogen rather than solely an opportunistic organism. Nonetheless, it is possible those cats may have had underlying conditions that were not recognized. Further testing of those enterococcal isolates for presence of virulence factors (such as extracellular surface protein or aggregation substance) may have provided more information regarding their pathogenicity; however, conclusive associations between presence of virulence factors and uropathogenicity of enterococci in dogs and cats have not yet been documented, this testing is not routinely performed on aerobic urine cultures at the Kansas State Veterinary Diagnostic Laboratory, and those isolates were no longer available at the time this retrospective study was prepared.¹⁴ Presence of bacteriuria (cocci) and pyuria were common in concurrent urine samples, giving support to enterococci causing infection and contributing to clinical signs; however, it is possible that WBCs were present for other reasons (urinary catheter, stones, etc.). Unfortunately, the small number of feline urine specimens with enterococcal growth included in the current analysis negatively affected the power, inhibiting the authors' ability to make statistical conclusions about the relationships between various culture, sediment, and clinical findings.

In dogs, clinical signs of lower urinary tract disease were observed in slightly fewer than 50% of all patients with a significant enterococcal UTI, and the majority of enterococcal UTIs were found in patients characterized as having complicated UTIs. Although no association could be identified between the presence of clinical signs and simple versus complicated UTIs, the low sample size limited the power to detect any association, and further research with increased urine specimens would be more conclusive. A clinician may submit a urine culture from either a dog or cat without clinical signs of lower urinary tract disease if the clinician is concerned about the consequences of failing to treat an overlooked UTI and/or if it is suspected that clinical signs may be present but not easily observed. Lack of clinical signs in some patients may be due to the opportunistic nature of this organism and raises the question of colonization versus true infection. In human healthcare, most patients with asymptomatic enterococcal bacteriuria are believed to be colonized rather than infected and do not require treatment.¹⁵ Elimination of predisposing factors, such as an indwelling urinary catheter, is recommended when possible and may eliminate enterococcal bacteriuria.¹⁵ The Infectious Disease Society of America has concluded that asymptomatic bacteriuria in most human patients is not harmful, leading to current recommendations that only pregnant women and individuals undergoing either prostate or invasive urogenital surgery should be screened with urine culture when asymptomatic.¹⁶ In veterinary medicine, the consequences of not treating patients with enterococcal bacteriuria have not been thoroughly evaluated, and elimination of predisposing conditions is not always possible. Until prospective studies are available to determine risk of consequences (i.e., pyelonephritis, bacteremia) in veterinary patients with enterococcal bacteriuria, global recommendations cannot be made and therapeutic decisions continue to be made on an individual basis considering concurrent conditions and perceived risk.

The majority (72.7% of feline and 74.2% of canine cases) of urine specimens in the current study with significant enterococcal growth were found to be Enterococcus-only specimens, rather than mixed with other bacterial species. Similarly, a 2008 study reported that 73.3% of canine urine samples that were positive for enterococcal growth had Enterococcus-only rather than mixed bacterial species.² In the current study, the presence of pyuria was associated with Enterococcus-only UTIs in dogs. Presence of WBCs in all dogs with an Enterococcus-only UTI supports the idea that inflammation and true infection may have been caused by Enterococcus spp. Based on anecdotal evidence, it has been recommended that when enterococci are found in combination with more pathogenic bacteria, such as Escherichia coli, the clinician should select an antimicrobial agent that targets the more pathogenic organism because it is believe that the enterococcal infection may resolve if the other organism is successfully eliminated.⁹ Prospective longitudinal research is needed to further investigate this recommendation. Until available, the CFUs/mL urine, enterococcal species and risk of ascending infection should be considered when making therapeutic decisions, and therapy should be targeted towards both organisms whenever possible.

Interpreting enterococcal susceptibility is challenging because CLSI breakpoints specific to canine and feline enterococcal UTIs are not available for any antimicrobial agent. Enterococci-specific breakpoints are available for ampicillin, penicillin, and erythromycin, but these are not canine, feline, or urinary specific.¹¹ Similarly, there are CLSI breakpoints for enrofloxacin that are canine and UTI specific, but not specific for enterococci.¹¹ Until the CLSI is able to provide more specific breakpoint recommendations, clinical laboratories use breakpoints suggested for other animal species or humans, other bacterial organisms (often Staphylococcus spp.), and often other body sites when interpreting enterococcal UTI isolates. That is a limitation for interpreting results of this study as well as interpreting clinical culture reports with enterococcal growth.

From a list of commonly used antibiotics for UTIs (amoxicillin or ampicillin, amoxicillin trihydrate/clavulanate potassium, cephalosporins, fluoroquinolones, chloramphenicol, trimethoprim-sulfa), clinicians treating enterococcal UTIs are limited because cephalosporins and trimethoprim-sulfa are known to have little clinical efficacy for most enterococci infections.^9,12,13 This study documented enterococcal resistance to enrofloxacin in E. faecalis (10 out of 24 canine isolates; 7 out of 11 feline isolates) and E. faecium (3 out of 4 canine isolates). The additional E. faecalis isolates with a MIC of 1 μg/mL (13 out of 24 canine isolates; 2 out of 11 feline isolates) may also not clinically respond to enrofloxacin because those would be in an intermediate susceptibility range. These results support the previous suggestion that enrofloxacin may have limited efficacy against enterococci.⁹ However, with the retrospective nature of this review it was not possible to obtain a full antibiotic history on these patients to determine if previous fluoroquinolone use could have influenced these results. Although E. faecium was rarely isolated in this study, E. faecium isolates showed a high proportion of antimicrobial resistance and were 4.5× more likely than E. faecalis to be MDR. These findings emphasize the need for culture and susceptibility testing to assist in antimicrobial selection rather than choosing empirical therapy for enterococcal infections.

Conclusion

Enterococci were routinely isolated from the urine of dogs and cats in quantities sufficient to be defined as UTIs. E. faecalis was the most common enterococcal species isolated from the urine of dogs and cats, followed by E. faecium from dogs. Most urine specimens with enterococcal growth contained enterococci alone, rather than a mixed bacterial population. Medical record review revealed that the majority of patients whose urine grew significant quantities of enterococci had complicated UTIs and fewer than 50% showed clinical signs of UTIs. E. faecium isolates had the highest proportion of antimicrobial resistance and were most likely to carry multidrug resistance.