Introduction

Urinary tract infections (UTI) occur commonly in dogs.¹ The etiology of UTI is usually multifactorial, and overall health, anatomy, and immune competence of the patient may all play a role.^2,3 UTI can be classified as simple, complicated, or recurrent.^3,4 Simple UTI (SUTI) are uncomplicated and occur sporadically in otherwise healthy animals with a normal urinary tract.⁴ Complicated UTI (CUTI) are diagnosed in animals with comorbidities such as diabetes mellitus, abnormal urinary tract conformation, or recurrent UTI that are characterized by the occurrence of >3 episodes in a 12 mo period.⁴ Most UTI are likely the result of an ascending infection, and enteric and cutaneous microorganisms are thought to be common sources of infection.^2,3,5–10

Although dogs with SUTI have often been empirically treated, current standards recommend antimicrobial susceptibility testing, as this allows for the selection of the optimal antimicrobial therapy and for adequately monitoring resolution of the UTI.⁴ Treatment failure can be the result of several factors including inappropriate antimicrobial selection, pharmacokinetic issues (e.g., biofilm formation, deep-seated infection), presence of a CUTI, and multidrug-resistant (MDR) bacteria.²

Several studies performed in veterinary teaching hospitals (VTHs) and referral centers reported the prevalence of bacterial species in cohorts of dogs with UTI.^5,7–10 Their stated objectives were to offer an in-depth analysis of selected data including causative bacteria from dogs with positive urine cultures, to identify bacterial resistance patterns and their changes over time, and to characterize persistent UTI and reinfections.^5,7–10 However, only few studies included detailed clinical data about the dogs in that clinical setting.^8,10

The present study provides detailed data on clinical parameters, urinalysis, and bacterial isolates from a cohort of dogs with positive urine culture seen at the Louisiana State University (LSU) VTH over a 3 yr period. Its purpose was to evaluate the spectrum of bacteria involved in UTI and their susceptibility to antimicrobials, including multidrug resistance, clinical parameters, and results of urine sediment analysis. Particular points of interest included evaluation of possible differences in bacterial isolates and their antimicrobial susceptibility in dogs from several clinically defined subgroups and the presence or absence of clinical signs of lower urinary tract disease (CS-LUTD) in this patient cohort.

Materials and Methods

Results

From January 1, 2012, to December 31, 2014, 1436 canine urine samples collected at the LSU VTH were submitted to the laboratory for aerobic culture and sensitivity tests, and 349 (24.3%) obtained from 288 dogs were positive. Of these 288 dogs, 208 satisfied the inclusion criteria for the study. Forty-eight dogs had more than one positive urine culture during the study period (median 2, range 2–10).

The most common breeds in the study population were Labrador retrievers (n = 20), mixed-breed dogs (n = 17), dachshunds (n = 14), miniature schnauzers (n = 12), and shih tzus (n = 10). There was a significant difference in the distribution of these 5 breeds between hospital and study population (P = .009). This was because of an overrepresentation of miniature schnauzers (5.8% versus 2.2%; P = .01). There were 142 females (123 spayed) and 66 males (51 neutered). When compared with the hospital population, females were markedly overrepresented among dogs with UTI (68.3% versus 51.2%; P < .0001). The median age of the studied cohort was 9 yr (range 0.25–18).

A total of 225 bacterial isolates were obtained from the 208 dogs at the time of initial urine culture; their distribution is shown in Table 1. Single infections were found in urine samples from 189 dogs (90.9%), whereas 19 (9.1%) had infections with multiple strains of bacteria. Among the most commonly isolated bacteria, the percentages of cultures revealing the presence of another bacteria (coinfections) were as follows: Klebsiella spp. (6/24; 25%), enterococci (7/29; 24%), staphylococci (5/26; 19%), and Escherichia coli (12/107; 11%). Susceptibility of bacterial isolates to amikacin, amoxicillin and clavulanic acid, ampicillin, cefovecin, cephalothin, chloramphenicol, enrofloxacin, and trimethoprim sulfamethoxazole is summarized in Table 2. Sixty-one strains (27.1%) were determined to be MDR. Among bacteria isolated from more than two samples, Enterobacter spp. had the highest rate of MDR with 6 of 9 (66.7%) and Streptococcus spp. the lowest with 0 of 9. The prevalence of MDR was 29% for E coli (31/107) and Staphylococcus spp. (9/31) isolates and 17.2% (5/29) for Enterococcus spp. isolates. Among staphylococci, 1 of 3 Staphylococcus aureus and 8 of 23 (34.8%) Staphylococcus pseudintermedius were methicillin resistant.

TABLE 1 Prevalence of Various Bacteria Among 208 Dogs with Urinary Tract Infections

View Fullscreen

TABLE 2 Susceptibility of Bacterial Isolates to Eight Antimicrobial Drugs Among Dogs with UTI, SUTI, CUTI, and PN; Among Dogs Presented for Primary Care or Referred Dogs; and Among Dogs Who Received Antimicrobial Treatment in the 30 Days Prior to Presentation

View Fullscreen

There were 23 dogs presented for primary care from whom 24 bacterial isolates were cultured, whereas 185 dogs were referred with 201 bacterial isolates. There was no statistically significant difference in bacterial distribution between the studied groups (Table 1). Also, no difference could be detected in antimicrobial susceptibility of bacterial isolates (Table 2). Although the prevalence of MDR in E coli appeared lower in primary cases (1/13 [7.7%] versus 30/94 [31.9%] in referred cases), this difference was not significant (P = .29). Similarly, there was no difference in MDR rate in staphylococci (1/1 in primary cases versus 8/30 [26.7%] in referred cases; P = .10).

SUTI were diagnosed in 41 dogs, whereas 157 dogs had CUTI and 10 had pyelonephritis. Identified comorbidities are listed in Table 3, and their overall prevalence was high (n = 165, 79.3%). It was higher in referred dogs (154/185, 83.2%) than in primary cases (10/23, 43.5%; P = .0004). SUTI were more common among primary cases (10/23, 43.5%) than among referred dogs (32/185, 17.3%; P = .01). There was no significant difference in the bacterial distribution (Table 1) or susceptibility to antimicrobials (Table 2) between dogs with SUTI or CUTI. There was no significant difference in the prevalence of MDR bacteria between dogs with SUTI, CUTI, and pyelonephritis, although both MDR E coli and staphylococci appeared less frequently in dogs with SUTI than in those with CUTI (MDR E coli 4/27 [14.8%], 26/86 [30.2%], and MDR staphylococci 1/16 [6.2%], 9/29 [31.4%] for SUTI and CUTI, respectively).

TABLE 3 Prevalence of Comorbidities and Recurrent UTI (Risk Factors) in 157 Dogs with CUTI, 63 Dogs with Clinical Signs of LUTD, and 145 Dogs Without Such Signs

View Fullscreen

Thirty-nine dogs with 43 bacterial isolates had been pretreated with antimicrobials. Drugs used included potentiated aminopenicillins (n = 13), fluoroquinolones (n = 9), first-generation cephalosporins (n = 7), aminopenicillins (n = 4), macrolides (n = 3), third-generation cephalosporins (n = 2), doxycycline (n = 2), 1 of each meropenem and nitrofurantoin, and 3 unknown (5 dogs received >1 antimicrobials). Pretreatment with these agents did not cause a shift in the prevalence of individual bacteria (data not shown). However, isolates from pretreated dogs were overall significantly less susceptible to enrofloxacin (Table 2). There was no significant change in the prevalence of MDR E coli (6/23 [26.1%] for pretreated dogs versus 25/84 [29.8%] for no pretreatment). The same was true for MDR staphylococci (3/5 [60%] versus 6/26 [23.1%]) or MDR enterococci (2/7 [28.6%] versus 3/22 [13.6%]), although they both appeared more prevalent in pretreated dogs.

In 15 dogs, urine cultures were submitted immediately after removal of an indwelling urinary catheter that had been in place for a mean duration of 6.1 days (range 3–10). Only 3 of them received antibiotics while they were catheterized. In these dogs, there was no difference in prevalence of various bacteria (data not shown), susceptibility to antimicrobials, or multidrug resistance (data not shown).

Only 63 dogs showed CS-LUTD (30.2%), whereas 162 (69.8%) did not show any such signs. There were no differences in distribution of bacterial species, sensitivity patterns of isolated bacteria, or prevalence of multidrug resistance bacteria between dogs with and without CS-LUTD. Prevalence of CS-LUTD was higher in primary cases (13/23, 56.5%) than in referred dogs (50/185, 27.0%; P = .007). Prevalence of CS-LUTD was not significantly different between dogs with SUTI (13/41, 31.7%) and with CUTI (51/157, 32.4%). However, CS-LUTD were not observed in any of the 10 dogs with pyelonephritis. Dogs with CS-LUTD were more likely to have anatomical defects, and dogs without documented CS-LUTD were more likely to have endocrinopathies or severely restricted mobility (Table 3).

Urine sediment analysis revealed pyuria in 145 dogs (69.7%) without significant difference in prevalence among dogs with (48/63, 76.2%) or without CS-LUTD (97/145, 66.9%). Hematuria was detected in 99 dogs (47.6%), and its prevalence was higher in dogs with CS-LUTD (39/63, 61.9%) than in those without CS-LUTD (60/145, 41.4%; P = .02). Pyuria and/or hematuria were present in 154 dogs (74.0%) with no significant difference between dogs with (49/63, 77.8%) and without CS-LUTD (105/145, 67.7%). Altogether, 40 of 208 (19%) dogs had no evidence of lower UTI based on clinical signs or urinalysis. Bacteriuria was present in 173 cases (83.2%) and was detected more frequently among dogs with pyuria (134/145, 92.4%) than among those without (39/63, 61.9%; P < .0002). It was also more commonly present in dogs with (90/99, 90.9%) than in dogs without hematuria (83/109, 76.1%; P = .01). However, there was no significant difference in prevalence of bacteriuria between dogs with and without CS-LUTD (55/63, 87.3% and 118/145, 81.4%, respectively). Finally, there was no difference in the prevalence of pyuria, hematuria, or bacteriuria if dogs had SUTI, CUTI, or pyelonephritis.

Discussion

The purpose of this retrospective study of a large group of dogs with positive urine cultures was to evaluate the spectrum of bacteria involved in UTI and their susceptibility to antimicrobials including multidrug resistance, clinical parameters, and results of urine sediment analysis.

The overall distribution of bacteria isolated from our canine cohort was similar to what has been previously reported.^5–10 There was no significant difference in the distribution of bacterial species among clinically defined subgroups. This is in accordance with other studies reporting 47% of isolates as E coli among dogs with persistent UTI, and 42–69% of urinary isolates in dogs with diabetes mellitus and/or hyperadrenocorticism.^10,14,15 It is, however, in contrast to a recent study that documented higher prevalence of E coli among dogs with SUTI (when compared with CUTI or pyelonephritis).⁸ In addition, no difference was observed in the distribution of bacterial isolates between dogs with and without documented CS-LUTD, similar to previous reports on subclinical CUTI.^14,15

Antimicrobial resistance can be constitutive or acquired following genetic mutation or plasmid transmission.¹⁶ We evaluated it in two ways. First, a clinical approach was used to determine what percentage of bacterial isolates would be susceptible to treatment with each antimicrobial of a panel of commonly tested drugs. With exception of ampicillin and cephalothin, isolates had a sensitivity of ≥70% to each of the other antimicrobials from the panel. Strikingly, urine bacteria from >83% dogs were susceptible to trimethoprim and sulfamethoxazole, comparable with a recent report.⁸ This proportion was even higher (93–96%) in dogs with SUTI and those presented for primary care, confirming the validity of current guidelines for treating UTI.⁴ Prior treatment with antimicrobials decreased the susceptibility of isolates to enrofloxacin, and the type of antimicrobial used did not seem to matter. This underscores the necessity to reculture the urine and reassess antimicrobial susceptibility in dogs who do not respond to first-line antimicrobials, as recommended in current guidelines, rather than empirically switch to a broader spectrum drug such as a fluoroquinolone.⁴ Antimicrobial resistance patterns may increase with indiscriminate use of antibiotics, although the evidence in veterinary medicine is still inconclusive.^16–18 Changes in antimicrobial susceptibility may also result from other factors such as intestinal colonization with resistant bacterial strains associated with changes in the intestinal microbiome, urinary colonization and infection with resistant bacteria during antibiotic treatment, or with resistant environmental bacteria in veterinary care facilities.^2–4

Second, we used relatively stringent, family-specific criteria taking into account each bacterial family’s intrinsic antimicrobial resistance pathways to define MDR.¹² A significant percentage of the most prevalent isolates (29% of E coli and staphylococci) were found to be MDR, slightly lower than in previous reports.^8,19 Underlying illness (a common problem in dogs from our cohort), prolonged hospitalization, surgical intervention, and prior antibiotic treatment with beta-lactams and/or fluoroquinolones were previously identified as potential risk factors for development of extraintestinal infections with MDR E coli and Enterobacter.²⁰ Unlike what was reported previously, the prevalence of MDR in our dogs was not higher with CUTI than with SUTI; however, both MDR E coli and staphylococci appeared less frequently in dogs with SUTI than in those with CUTI.⁸ Similarly, although there was no significant difference, the prevalence of MDR staphylococci and enterococci appeared higher in dogs pretreated with antimicrobials. Multidrug resistance did not vary either in comparisons between any of the other clinically defined groups. It may be that, although resistance to specific antimicrobials increases among bacteria, this does not impact the onset of multidrug resistance. Alternately, the relatively small size of the clinical subgroups may have rendered our study underpowered to identify such differences.

The signalment of dogs in our cohort was similar to previous reports with an overrepresentation of female dogs and a mean age of 9 yr.^5,8 Miniature schnauzers were overrepresented, as had been reported before in another study, which stated that the odds ratio for this breed to have a positive urine culture was nearly 12, when compared with Jack Russell terriers, who had the lowest incidence of UTI in that cohort.⁹ Breed variation in anatomy, biology, and susceptibility to diseases that predispose to UTI were suspected to play a role, even though the exact reasons remain unclear. Interestingly, of the 12 miniature schnauzers included in our study, all of whom were referred, 11 had CUTI, 1 had pyelonephritis, and 7 had no CS-LUTD.

Most of the dogs included in this study had been referred, and based on observations made in comparable studies, a high proportion of dogs with CUTI is expected among dogs seen at a referral center.^8,10 The list of comorbidities and their order of frequency was also in line with other studies, with immunosuppression being the most common.^8,10 This is not surprising because long-term immunosuppression with glucocorticoids and/or cyclosporine was previously associated with UTI in 18–30% of dogs.^21,22 As reported in previous studies, severely restricted mobility, endocrinopathies, and kidney diseases were the next-most-common comorbidities.^8,10 Severely mobility-restricted dogs included animals with neurological disorders as well as dogs with pelvic fractures and other severe orthopedic conditions who were not able to void urine properly. This might have resulted in abnormal bladder emptying and urine retention and necessitated manual emptying or placement of a urinary catheter, and these procedures are known to increase the risk of UTI.^23,24 Dogs with diabetes mellitus or hyperadrenocorticism may be more susceptible to UTI as a result of decreased urine osmolality, glycosuria, decreased neutrophil function associated with diabetes mellitus, or suppression of the immune response secondary to hypercortisolemia.¹⁴ Interestingly, severely restricted mobility and endocrinopathies were significantly overrepresented among dogs without documented clinical signs of lower UTI. Both these comorbidities have been previously associated with subclinical bacteriuria (SCB).^{14,15,23–25} Dogs with severely restricted mobility may lack the ability to display classical signs of a UTI such as pollakiuria or dysuria.²⁵ Owners of dogs with endocrinopathies and polyuria/polydipsia may overlook signs of UTI; additionally, compromised neutrophil function may also lessen the overt signs of inflammation.^14,15 Finally, dogs with chronic kidney disease also have decreased urine osmolality, and increased risk of developing infections has been reported in humans with predialysis chronic kidney disease.²⁶

Anatomical defects were not as common as in other studies; this may represent a true variation in our cohort or reflect the fact that clinicians’ interpretation of abnormalities such as recessed vulva was more restrictive than in other studies.^8,10 Interestingly, anatomical defects were more likely to occur in dogs showing signs of lower UTI. It may be that the presence of clinical signs of lower UTI encouraged clinicians to look for anatomical defects such as a hooded vulva.

The low prevalence of documented CS-LUTD in our cohort was striking. In previous studies, they were detected at presentation in only <5–46% of dogs with positive urine cultures.^10,14,15 The high prevalence of SCB in dogs with endocrinopathies, chronically paralyzed dogs, or dogs receiving cyclosporine or long-term steroid treatment was also described in previous reports.^{14,15,21,22,25} Owners may not observe their dogs closely enough to detect and report subtle changes in urination behavior or urinary incontinence.³ Therefore, it is possible that the prevalence of SCB in this study is exaggerated. In the absence of associated clinical signs, it is unknown which dogs with positive urine culture would benefit from antimicrobial treatment.^3,4 Under these circumstances, detection of pyuria and/or hematuria could be considered an additional help in the screening and ultimately in the decision to treat dogs with antimicrobials.³ However, in humans, pyuria is frequently observed with asymptomatic bacteriuria, and its presence is not considered an added risk factor.²⁷ In our cohort, two-thirds of dogs with SCB had pyuria or hematuria in their urine sediment. Hematuria was more commonly detected in dogs with CS-LUTD than in those with SCB. In the end, 19% of all dogs with positive urine cultures had neither documented clinical signs of lower UTI nor evidence of inflammation in their urine sediment. It is possible that some of these dogs may have had pyelonephritis that was not recognized by the clinicians because the disease is known to represent a diagnostic challenge.¹³ There is a need for future research evaluating the benefits of antimicrobial treatment in dogs with SCB.

Limitations of this study included its retrospective design, which relies entirely on the attending clinicians for making decisions in the diagnostic approach and management of patients. Moreover, although the total numbers of dogs and bacterial isolates included were appropriate, several of the clinically defined subgroups were small, and comparison of these groups to detect differences in proportions may have lacked statistical power. In addition, antimicrobial sensitivity patterns were determined using in vitro DD and not minimal inhibitory concentrations. Antimicrobials with high concentration of intact drug or active metabolites in the urine may have clinical efficacy in spite of a lack of in vitro effect.⁴ Although limited CLSI-approved urine-specific minimum inhibitory concentrations breakpoints are available, there are none when using DD. However, current guidelines recommend using serum breakpoint when interpreting susceptibility of urinary isolates.⁴ The interpretation and applicability of the DD susceptibility patterns to clinical situations is subject to caution, even though the method is still considered appropriate in veterinary medicine.^11,28 Finally, this data originated from a VTH that sees a majority of referred cases and only a small proportion of primary cases. Therefore, these results may not fully apply to primary care veterinary facilities.

Conclusion

Nearly 80% of dogs in this study had one or more risk factors (comorbidity or recurrent UTI). Interestingly, only 30% of dogs with positive urine culture were reported to show signs such as pollakiuria, dysuria, stranguria, and macroscopic hematuria. This highlights the importance of performing urinalysis and urine culture in dogs with diseases commonly associated with UTI and to carefully consider the benefits of antimicrobial treatment on a case-by-case basis. The distribution of bacterial isolates in our cohort was comparable with previous studies. Prevalence of MDR bacteria was moderate, and no significant changes were observed across all the clinical subgroups. Bacterial isolates from dogs treated with antimicrobial agents in the month preceding presentation showed decreased susceptibility to enrofloxacin, thus underscoring again the absolute necessity of urine culture and sensitivity to facilitate judicious and discriminate use of antimicrobials for proper management of each dog.