Introduction

Enterococcus species are important agents of bacterial cystitis and subclinical bacteriuria (SBU) in dogs and are often the second or third most common isolate cultured, accounting for 8–27% of isolates.^1–4 A previous study in dogs identified risk factors for enterococcal bacteriuria, relative to Escherichia coli bacteriuria, as the presence of a lower urinary tract anatomic abnormality (i.e., recessed vulva), urolithiasis, or lower urinary tract neoplasia.² This study and others investigating Enterococcus spp. bacteriuria in humans have suggested that Enterococcus spp. may colonize the urinary bladder when local or systemic host defenses are diminished.^5,6

Previous canine studies have identified Enterococcus spp. as common coisolates in polymicrobial urinary tract infection (UTI) often with other uropathogens such as E. coli and Proteus mirabilis.^1,7–9 A recent study compared polymicrobial UTI with E coli and Enterococcus spp. with monomicrobial growth of each organism. This study found that compared with monomicrobial enterococcal bacteriuria, polymicrobial infections were more commonly implicated in recurrent UTI.⁹ The prevalence of polymicrobial UTIs in dogs has been previously reported as 13–23%.^1,4,7–9 In these clinical scenarios, it is unclear if, or to what degree, the Enterococcus isolate contributes to clinical signs when compared with the other organism. Anecdotally, some clinicians will target treatment of the nonenterococcal organism with an assumption its treatment may lead to resolution of clinical signs or bacteriuria. Challenges associated with treatment of Enterococcus spp. include their intrinsic resistance to many empiric antimicrobials, as well as their ability to acquire resistance in the face of antimicrobial therapy.^10,11

Previous studies have evaluated Enterococcus spp. bacteriuria compared with E coli bacteriuria, but to the authors’ knowledge, no studies have investigated Enterococcus spp. bacteria compared with all other agents of bacteriuria.² Additionally, few studies have reported on the outcomes of antimicrobial therapy. Thus, the primary goals of this study were to evaluate whether presence of lower urinary tract clinical signs was more commonly associated with enterococcal bacteriuria compared with other pathogens, whether enterococcal bacteriuria is more common in dogs with systemic comorbidities compared with other urinary pathogens, and whether resolution of clinical signs with empiric antimicrobial therapy is different in dogs with Enterococcus spp. bacteria compared with all other agents of bacteriuria. A secondary goal was to determine the frequency at which Enterococcus spp. are detected in polymicrobial UTIs in dogs, compared with other bacteria. We hypothesized that, compared with other bacteria, enterococci would be more commonly isolated in dogs with SBU and polymicrobial infections, and that when local or systemic comorbidities were present, enterococcal bacteriuria would be more common. We also hypothesized that dogs with nonenterococcal bacteriuria would have resolution of clinical signs more frequently than dogs with enterococcal bacteriuria with antimicrobial therapy.

Materials and Methods

A targeted search of clinical microbiology laboratory records was performed retrospectively to identify all positive canine urine cultures from April 2017 to December 2019. The clinical microbiology laboratory serves all patients at the Matthew J. Ryan Veterinary Hospital of the University of Pennsylvania, including primary care patients as well as tertiary referral patients. Data were subsequently filtered to identify urine specimens harboring Enterococcus spp. A single American College of Veterinary Medicine Board–certified (Small Animal Internal Medicine) author (E.J.S.) reviewed all medical records. A case was defined as a dog with a positive culture yielding Enterococcus at a clinically significant concentration (≥10³ colony forming units [CFU] per milliliter for urine collected via cystocentesis, ≥10⁴ CFU/mL for urine collected via catheterization, and ≥10⁵ CFU/mL for urine collected via free-catch).¹² Dogs with unavailable medical records were excluded from the study.

Data collected from medical records included signalment, prior medical history, and body weight. Physical examination (including body condition score and anatomic urinary abnormalities), clinical signs (including hematuria, pollakiuria, stranguria, and/or dysuria), and urinalysis findings were reviewed and included if available. When available, antimicrobial therapy and resolution of clinical signs within 2 wk was evaluated. All treatments complied with the contemporary standard of care. Empiric antimicrobial choices were made by the primary clinician based on International Society for Companion Animal Infectious Diseases guidelines or based on results of culture and susceptibility. Records were evaluated for a historical or current diagnosis of endocrinopathies, myelopathies, gastrointestinal disease, renal disease based on International Renal Interest Society guidelines, immunosuppressive medication administration, urinary calculi, or urinary tract neoplasia. International Society for Companion Animal Infectious Diseases definitions for sporadic bacterial cystitis (first instance of bacteriuria within 6 mo) and recurrent bacterial cystitis (more than one instance in the preceding 6 mo) were used to categorize dogs based on timing of bacteriuria. Records were also evaluated for previously diagnosed urinary tract abnormalities not readily evident on physical examination, including ectopic ureters and persistent paramesonephric septal remnant.

Controls were selected from the same database containing all positive aerobic urine cultures. Controls were matched to cases based on the year in which the culture was performed, sex, age, and weight. If more than three controls were available, the first three were chosen alphabetically. Cultures were excluded from analysis if they did not meet the definition for significant growth defined for the Enterococcus spp. cases. For every aerobic urine culture positive for Enterococcus spp. growth, three aerobic urine cultures positive for any other bacterial species were selected as controls. Data retrieved from control medical records were the same as for cases.

For complete urinalysis, samples were allowed to warm to room temperature and were then analyzed within 1 hr. Urine specific gravity was determined using a refractometer^a. Dipstick analysis^b was followed by urine sediment evaluation using 3–5 mL of centrifuged urine with a minimum requirement of 1 mL. Following centrifugation, urine supernatant was decanted and one drop of resuspended urine sediment was pipetted onto a microscope slide with a coverslip and allowed to settle for 1 min. Urine sediment was evaluated microscopically using established grading criteria for formed elements, at both ×10 (low-power fields) and ×40 (high-power fields) magnification. When bacteriuria was suspected, a dried sediment or cytospin slide was prepared and stained with Wright Giemsa to confirm the presence of bacteria.

Samples submitted to the clinical microbiology laboratory were required to be submitted in a sterile container and submitted to the laboratory or refrigerated within 2 hr of collection. Samples were not stored for longer than 36 hr before submission. A calibrated 1 µL loop was used to streak urine specimens in a quantitative fashion to tryptic soy agar with 5% sheep’s blood, MacConkey, and colistin-nalidixic acid agar plates^c. Plates were incubated under ambient atmospheric conditions for 16–24 hr at 35 ± 2°C. Cultures were assessed for bacterial growth by trained technical staff and bacterial concentrations recorded as one of the following four categories: ≤10,000 CFU/mL; >10,000 to <50,000 CFU/mL; ≥50,000 to <100,000 CFU/mL; and ≥ 100,000 CFU/mL. Following subculture, pure isolates were identified and antimicrobial susceptibility testing (AST) performed using an automated system^d. At the time of the study, no veterinary AST breakpoints have been published for enterococci; therefore, AST interpretations were made using human clinical breakpoints, which have been included in the Clinical Laboratory and Standards Institute Vet01S document.¹³ The laboratory only routinely reports penicillin and nitrofurantoin results to clinicians given these limitations.

Results

A total of 641 aerobic cultures were positive for growth during the study period, including 97 cases with significant enterococcal growth. One case was excluded because of an unavailable medical record, leaving 96 cases included in analysis. There were 288 year-, age-, and sex-matched control cases selected from aerobic urine cultures positive for organisms other than Enterococcus spp. All of the variables collected from the records were analyzed.

The demographic data for cases and controls are presented in Table 1. The five most common breeds represented in the case population were mixed-breed dog (n = 28, 29%), pug (n = 15, 16%), bearded collie (n = 7, 7%), Labrador retriever (n = 5, 5%), and Saint Bernard (n = 5, 5%). In the control group, the three most common breeds were mixed-breed dog (n = 97, 34%), Labrador retriever (n = 22, 8%), and American pit bull terrier (n = 15, 5.2%). Compared with mixed-breed dogs, pugs, bearded collies, and Saint Bernards had higher odds of having enterococcal bacteriuria than another bacterial infection (odds ratio [OR] 7.4, 95% confidence interval [CI] 2.6–19.9, P ≤ .001; OR 24.3, 95% CI 2.9–205.5, P = .003; OR 17.3, 95% CI 1.9–154.4, P = .01, respectively). Other breeds did not have a higher risk of developing Enterococcus spp. bacteriuria.

TABLE 1 Demographic Data of Cases and Controls

View Fullscreen

Comorbidities found to have a significant correlation with enterococcal bacteriuria are summarized in Table 2. No statistically significant difference was identified in other comorbidities evaluated, administration of corticosteroids or other immunosuppressive therapy, lower urinary tract neoplasia, or a known history of urinary calculi.

TABLE 2 Comorbidities Identified in Dogs with Enterococcus spp. Bacteriuria and Dogs with Bacteriuria of Another Organism

View Fullscreen

In 204/384 (53%) dogs, this was the first instance of bacteriuria in 6 mo, diagnosed in significantly fewer case dogs (31/96, 32.3%) compared with control dogs (173/288, 60.1%, P ≤ .001). Lower urinary tract signs were significantly more common in controls (114/288, 40%) compared with cases (26/96, 27%, P = .02). Specifically, control dogs (64/288, 22.2%) were significantly more likely to have gross hematuria than case dogs (9/96, 9.4%, P = .01). Urinary incontinence was significantly more common in cases (43/96, 45%) than controls (85/288, 30%; P = .01).

In the combined population of 384 case and control dogs, a total of 339 urinalyses (urine dipstick and/or urine sediment analysis) were performed (86.5%). Samples were collected via cystocentesis (202/339, 59%), free-catch (77/339, 22.5%), and urinary catheterization (29/339, 8.5%), at the time of cystoscopy (1/339, 0.3%), and by other methodologies including collection from a subcutaneous ureteral bypass system or at the time of surgery (14/339, 4.1%). The collection method was not stated for 19 cases (5.6%). Collection methods were not significantly different between cases and controls. More controls (211/270, 78%) than cases (59/81, 73%) had dipstick hematuria (P = .02). Other urinalysis findings including urine specific gravity, pH, proteinuria, pyuria, and the presence of bacteria on sediment examination were not significantly different between cases and controls.

Isolates from the Enterococcus spp. cases were Enterococcus faecalis (68/96, 71%), Enterococcus faecium (21/96, 22%), Enterococcus durans/hirae (2/96, 2%), Enterococcus gallinarium (3/96, 3%), and Enterococcus casseliflavus (2/96, 2%). Sixteen organisms were isolated from the control group including E coli (166/288, 57.6%), Staphylococcus pseudintermedius (47/288, 16.3%), Proteus mirabilis (41/288, 14.2%), Klebsiella pneumoniae (15/288, 5.2%), Pseudomonas aeruginosa (4/288, 1.4%), Citrobacter braakii (3/288, 1%), Klebsiella oxytoca (3/288, 1%), and one (.35%) each of the following: Staphylococcus schleiferi, Pasturella canis, Serratia marcescens, Morganella morganii, Staphylococcus capitis, Staphylococcus hominis, Staphylococcus haemolyticus, Klebsiella aerogenes, and Enterobacter cancerogenus. Most urine cultures from the entire population showed monomicrobial growth (329/384, 86%). Table 3 reports the breakdown of polymicrobial coisolates. From the case population of 96 cultures, 50/96 were monomicrobial (52.1%), 42/96 had two bacterial isolates (43.8%), and 4/96 had three bacterial isolates (4.1%). The control population with 288 urine cultures had 279/288 monomicrobial infections (96.9%) and 9/288 had two bacterial isolates (3.1%). Although the majority of dogs had monomicrobial infections, cases were significantly more likely to have a polymicrobial infection than controls (OR 28.52, 95% CI 12.63–69.62, P ≤ .001).

TABLE 3 Distribution of Bacteria in Enterococcal Cultures with Two Microbial Isolates

View Fullscreen

Although follow-up data were limited, this study did evaluate the resolution of clinical signs (lower urinary tract signs or urinary incontinence) within 2 wk of starting antimicrobial therapy as reported by the owner. The case group had 69 dogs with clinical signs and the control group had 199 dogs with clinical signs. Follow-up information was available for 36 cases and 155 controls. In the case group, 30/36 (83%) received antimicrobial therapy, and 16/36 (44%) showed resolution of clinical signs within 2 wk. For symptomatic dogs with enterococcal bacteriuria, there was no association between resolution of clinical signs and administration of antimicrobials (P = .19). In the control group, 152/155 (98%) received antimicrobial therapy, and 123/155 (79%) showed resolution of clinical signs within 2 wk. In the control population, there was an association between resolution of clinical signs and administration of antimicrobials (P = .01).

Discussion

This retrospective study contributes to the growing body of literature about the role of Enterococcus spp. as uropathogens as well as agents of SBU and polymicrobial bacteriuria. This study identified breed-associated risk of having enterococcal bacteriuria compared with other bacteriuria. Recurrent UTI, urinary incontinence, recessed vulva, myelopathy, renal disease, and hyperadrenocorticism were more common in patients with enterococcal bacteriuria. Unlike previous studies, this study did not find an association of enterococcal bacteriuria and the presence of lower urinary tract neoplasia or urolithiasis. This study also found that more dogs with other bacteriuria responded to empiric antimicrobial therapy than dogs with enterococcal bacteriuria.

Pugs, bearded collies, and Saint Bernards had higher odds of developing enterococcal bacteriuria compared with another bacterial infection when compared with mixed-breed dogs. This is a new finding, as none of these breeds were overrepresented in prior studies. These breeds are more likely to have comorbidities significantly associated with enterococcal UTIs. Pugs disproportionately develop certain spinal diseases, including congenital abnormalities and various degenerative processes.¹⁴ Given the increased odds of an enterococcal UTI in dogs with myelopathies when compared with dogs without neurological disease, this may explain the increased odds of enterococcal bacteriuria in this breed. Bearded collies have an increased risk of urinary incontinence when compared with other breeds, and this may explain their predilection for enterococcal UTIs.¹⁵ In a previous study, Saint Bernards had an increased risk of bacterial cystitis progressing to a recurrent or persistent UTI, and at a younger age than other breeds.³ These dogs are also at an increased risk for orthopedic disease, which may contribute to incomplete bladder emptying.¹⁶

Like previous reports, our study found an increased risk of enterococcal bacteriuria in dogs with a recessed vulva and myelopathy. A recessed or hooded vulva is a known risk factor for recurrent UTI owing to anatomic distortion of the lower urinary tract.⁴ Dogs with myelopathies can have abnormal micturition and may be unable to void their bladders completely, leading to urine retention and increased risk of infection. Dogs with a myelopathy might not demonstrate classic lower urinary tract signs given their neurologic dysfunction, and the presence of lower urinary tract signs can be underestimated in this population.⁷ It is possible that Enterococcus spp. are a primary uropathogen or have colonized the urinary bladder in these dogs owing to recurrent infections with other bacterial pathogens or previous antibiotic administration. Prior antimicrobial administration was not assessed in this study and could also influence the bacterial population.

More dogs with renal disease and hyperadrenocorticism had Enterococcus spp. bacteriuria compared with controls. Previous veterinary studies suggested chronic kidney disease (CKD) as a risk factor for enterococcal bacteriuria but were underpowered to detect a statistical difference compared with other bacterial isolates.¹ A recent study evaluated SBU, bacterial cystitis, and pyelonephritis in dogs with CKD. The incidence of a positive urine culture was 18%, and those cases were retrospectively diagnosed with SBU (45%), pyelonephritis (40%) or bacterial cystitis (6%) with no association between diagnosis and International Renal Interest Society CKD stage.¹⁷ A similar study in dogs with hyperadrenocorticism found that 46% of dogs with hyperadrenocorticism had bacteriuria and that less than 5% had lower urinary signs.¹⁸ In both studies, the numbers of dogs with Enterococcus spp. Growth were low, at one and three dogs, respectively. Renal disease is hypothesized to be associated with urinary tract infection because of decreased local defense mechanisms such as decreased urine concentrating ability.¹⁹ Hyperadrenocorticism is thought to increase risk of UTI due to immunosuppression from excess glucocorticoid, as well as through production of dilute urine.²⁰ Given these known risks, it is possible that more dogs with CKD and hyperadrenocorticism received urine cultures as part of a diagnostic evaluation; thus, more SBU was identified in both cases and controls with these comorbidities. Further studies are needed to clarify the association between Enterococcus spp. bacteriuria and these comorbidities in dogs and whether treatment of SBU should be considered in this population.

The lack of association between Enterococcus spp. bacteriuria with urolithiasis and urinary tract neoplasia in this study was an unexpected finding. Urolithiasis and urinary tract neoplasia predispose dogs to recurrent bacteriuria.¹² In humans, urinary catheterization is associated with enterococcal bacteriuria, which is suspected to be due to urothelial injury predisposing to infection.^5,6 Wood et al. found urolithiasis and lower urinary tract neoplasia more commonly in dogs with Enterococcus spp. bacteriuria compared with E coli bacteriuria, hypothesizing that these underlying conditions cause similar uroepithelial damage to indwelling urinary catheterization.² Infection-associated lower urinary tract stones are associated with urease-producing bacteria such as Staphylococcus spp. and Proteus spp.²¹ It is possible that the inclusion of other urease-producing bacteria in our control population led to an increased prevalence of urolithiasis in the control group and could explain the discrepancy in these two studies. It is also possible that more urine cultures are performed in patients with these comorbidities at different institutions.

In our study, the odds of recurrent bacteriuria were higher in Enterococcus spp. cases compared with controls. This is in agreement with multiple other studies that showed Enterococcus sp. as a common agent in recurrent urinary tract infection, either alone or as a polymicrobial infection.^2,4,9 The hypotheses in these and other human studies are that enterococcal bacteria may alter the local urinary bladder environment by suppressing the innate immune response, mediating adherence to fibrinogen and collagen to facilitate biofilm formation, and acidifying urine through lactic acid production.¹⁰ These functions may contribute to the increased risk of recurrent UTI as well as polymicrobial UTI with enterococcal bacteriuria. It is also likely that in our patient population, decreased local defense mechanisms increased the risk for recurrent and polymicrobial UTI. Enterococcal UTIs are less likely to respond to empiric antimicrobials given their intrinsic resistance, which could contribute to recurrent bacteriuria if urine cultures are not performed to identify the organism.

In this study, the odds of exhibiting clinical signs were lower with Enterococcus spp. bacteriuria compared with bacteriuria with other bacterial organisms, with the exception of urinary incontinence. Urinary incontinence could be a sign of a urinary tract infection or a sign of an underlying condition such as urethral sphincter mechanism incompetence, ectopic ureters, or other anatomic abnormalities.²² Given our other findings, it is more likely that the increased odds of urinary incontinence in the enterococcal bacteriuria group are related to underlying conditions such as a recessed vulva, myelopathy, or urethral sphincter mechanism incompetence rather than a sign of a lower urinary tract infection. Previous studies have investigated SBU in healthy dogs and dogs with comorbidities including diabetes mellitus, renal disease, obesity, and myelopathies, as well as in those receiving certain medications (oclacitinib). The organisms commonly implicated in these studies include the same causative organisms of bacterial cystitis including E. coli, Klebsiella spp., Staphylococcus pseudintermedius and Enterococcus spp.^{8,17–19,23,24} The previous studies investigating enterococcal bacteriuria found SBU rates of 44.8–56%, which were lower than the rate of 73% reported in the current study. The high rates of Enterococcus spp. in SBU contribute to the hypothesis that Enterococcus spp. may colonize rather than infect the lower urinary tract in some dogs.

The odds of polymicrobial growth on aerobic culture were higher for Enterococcus spp. than with other bacterial species. Previous human and veterinary studies have identified Enterococcus spp. as a common coisolate in polymicrobial urinary tract infections.^5,9,10 Although there is evidence to suggest that Enterococcus spp. may act as a primary pathogen in the urinary tract, the presence of other bacteria causing a UTI and secondary urinary tract inflammation may increase the risk for colonization with Enterococcus spp. as a secondary pathogen.³ In humans, Enterococcus spp. is also shown to cause local, innate immune suppression, which can allow for proliferation of less virulent bacteria.¹⁰ In our study, the majority of polymicrobial growth cases were associated with clinical canine uropathogens (predominantly E coli). Persistent colonization evidenced by subsequent positive bacterial cultures with the same organism was not investigated in this study. Anecdotally, some clinicians leave Enterococcus spp. untreated in a polymicrobial infection, with the assumption that treating the more pathogenic bacteria may lead to spontaneous resolution of the Enterococcus spp. infection and resolution of the patient’s clinical signs.¹¹ In this and other studies, Enterococcus bacteriuria caused SBU in many, but not all, cases, which calls the previously described practice to question. Further studies are needed to determine whether enterococcal SBU left untreated may lead to negative sequelae.

Although this study was not designed to investigate antimicrobial therapy, symptomatic control dogs had an increased odds of resolution of clinical signs after starting antimicrobial therapy compared with symptomatic Enterococcus spp. dogs. When Enterococcus spp. infections cause clinical signs, their inherent resistance to commonly chosen empiric antimicrobials (with the exception of amoxicillin) might lead to a lack of response.¹¹ This study did not investigate the relationship between the antimicrobial susceptibility and empiric antimicrobial chosen to treat symptomatic patients, although this scenario can be representative of clinical practice, where the clinician must make a choice about antimicrobial therapy before availability of culture results, which may take several days. This study did not account for other factors when evaluating resolution of clinical signs, but the presence of other diseases that are associated with lower urinary tract inflammation, including urolithiasis and urogenital neoplasia, were not statistically different between groups. The study was also not designed to evaluate follow-up culture information. Further prospective studies should focus on follow-up of clinical resolution in conjunction with microbiological resolution after proper antimicrobial therapy to determine whether targeting therapy toward the more pathogenic organism is best practice.

Limitations of this study are its retrospective nature and the inability to standardize and assess treatment and outcome. In an attempt to represent a hospital population accurately and increase the generalizability of the findings, this study had broad inclusion criteria and minimal exclusion criteria. A major limitation is that this study did not investigate prior antimicrobial use as a risk factor for development of enterococcal bacteriuria. Although the Microbiology Laboratory services both primary and tertiary care patients, the numbers of each were not reported here, and the tertiary referral patients may represent more complicated or severe cases. Some of the study findings, such as a lack of a difference in comorbidities between cases and controls, might be due to a small sample size and a type II statistical error. Correction for multiple testing was not performed because in early exploratory studies such as this one, the benefit of initial discoveries that can pave the way for future studies outweighs the risk of a type I statistical error in which a false-positive finding is reported.²⁵ The ORs presented here are crude ORs that might not hold up in multivariable logistic regression when adjustments are made for other covariates. Additionally, the CIs for some OR estimates are wide, likely reflecting the small numbers of dogs that fall into these categories. The results of this study should be viewed as preliminary and must be confirmed with larger studies that would allow for correction for multiple testing and examination of interactions between variables.

Conclusion

This study highlights new information about breed risk and the importance of Enterococcus spp. in polymicrobial urinary tract infections. Enterococcus spp. are important pathogens in dogs with recurrent bacteriuria, recessed vulva, myelopathy, CKD, and hyperadrenocorticism. This study confirms that Enterococcus spp. are important agents of SBU in dogs. There are clinically relevant differences in the findings of dogs with enterococcal versus other bacteriuria.