The urinary tract has a great capacity to avoid infection

Save this PDF as:

Size: px
Start display at page:

Download "The urinary tract has a great capacity to avoid infection"


1 J Vet Intern Med 003;17:6 631 Persistent Urinary Tract Infections and Reinfections in 100 Dogs ( ) M. Alexis Seguin, Shelly L. Vaden, Craig Altier, Elizabeth Stone, and Jay F. Levine A retrospective study was performed of 100 dogs with persistent urinary tract infections (UTIs) or reinfections presenting to the North Carolina State University (Raleigh, NC) Veterinary Teaching Hospital between 1989 and Criteria for selection included positive urine cultures within a 6-month period. Signalment, presence of predisposing disorders, urinalysis and urine culture results, and treatment strategies were extracted from the medical records. Dogs were a median age of 7 years when the UTI was 1st diagnosed. Dogs younger than 3 and older than 10 years were at increased and decreased risks, respectively, for reinfections or persistent UTIs. Spayed females were more common in the UTI population. More than half of the dogs were asymptomatic for a UTI at 1st presentation. Urine sediment examinations identified hematuria, pyuria, and bacteriuria in 47, 7, and 85% of the samples, respectively. The most commonly isolated organisms were Escherichia coli and Streptococcus/Enterococcus spp.; multiple isolates also were common. Of the isolates, 9.5% were resistant to achievable serum concentrations of all antibiotics commonly prescribed for PO administration. Dogs with abnormal micturition were more likely to have infections by organisms resistant to commonly prescribed antibiotics. Potentially predisposing disorders were identified in 71 dogs. A correction of these disorders was accomplished in 35% of these 71 dogs. Dogs given standard antibiotic therapy without addressing predisposing disorders experienced poor control of their UTIs; 74.5% of these dogs had an apparent disease-free interval (ADFI) of 8 weeks. By comparison, dogs in which predisposing disorders were corrected or those that were treated with low-dose, long-term antibiotic regimens subjectively had better control. Key words: Canine; Chronic urinary tract infection; Complicated urinary tract infection; Predisposing condition; Recurrent urinary tract infection; Resistance. The urinary tract has a great capacity to avoid infection despite its proximity to the anus and the potential for fecal contamination. Normal host defenses hypothesized to be important in preventing urinary tract infections (UTIs) include normal micturition, mucosal defense barriers, antibacterial properties of urine, specific anatomic structures, and systemic immunocompetence. 1, This capacity is shown in studies in which Escherichia coli either was inoculated once 3 or infused at a constant rate over a period of several weeks 4 into the bladder of healthy dogs. After inoculation was completed, dogs cleared the infection without being given antibiotics within 3 (1-time inoculation) to 5 days (constant-rate infusion). A UTI is unlikely in the presence of normal host defenses. 1,3,4 Despite these defense mechanisms, some dogs present with several positive urine cultures within a relatively short time. Multiple diagnoses of UTIs in these dogs may reflect From the Mesa Veterinary Hospital, Mesa, AZ (Seguin); and the Department of Clinical Sciences (Vaden, Stone), the Department of Microbiology, Pathology, and Parasitology (Altier), and the Department of Farm Animal Health and Resource Management (Levine), College of Veterinary Medicine, North Carolina State University, Raleigh, NC. Preliminary results from this study were presented orally at the 18th Annual Veterinary Medical Forum (ACVIM), Seattle, WA, 000. The abstract for this publication was published as follows: Seguin MA, Vaden S, Levine J, Stone E. Complicated urinary tract infections in 100 dogs: A retrospective study. J Vet Intern Med 000; 14:354. This work was completed at the College of Veterinary Medicine, North Carolina State University. Reprint requests: Shelly L. Vaden, DVM, PhD, Diplomate ACVIM, Associate Professor, Internal Medicine, Department of Clinical Sciences, College of Veterinary Medicine, North Carolina State University, 4700 Hillsborough Street, Raleigh, NC 7606; shelly Submitted March 11, 00; Revised September 17 and December 6, 00; Accepted February 5, 003. Copyright 003 by the American College of Veterinary Internal Medicine /03/ /$3.00/0 reinfection or failure to clear the initial infection because of a persistent or relapsing infection. 1,,5 Persistent or relapsing infections occur when the original infection is not fully resolved despite therapy. 1,5,6 Reinfection occurs when a new bacterial species or strain infects a patient after a period of urine sterility, which is documented by a negative culture. 1,5,7 Superinfections are variants of reinfections or persistent infections in which a different organism colonizes the urinary tract during the course of treatment for the original infection. 1,5 Reinfections and persistent UTIs for which appropriate treatment has been administered are likely to reflect abnormal host defenses. The alterations in host defenses may be intrinsic to the patient (eg, diabetes mellitus) or iatrogenic (eg, corticosteroid administration). Only a few of the many conditions suspected to predispose to UTIs have been studied A recent retrospective study characterized urine culture results from dogs with reinfections and persistent infections and attempted to identify associations with signalment, but it did not evaluate altered host defenses. 6 This retrospective study was conducted to characterize persistent UTIs and reinfections in 100 dogs evaluated at a veterinary teaching hospital. These dogs were evaluated for the presence of disorders potentially associated with increased risk for complicated infections. In addition, clinical presentation, urinalysis results, and bacterial isolates and their antimicrobial susceptibility patterns were recorded. The choices and relative success of different treatment strategies used in the management of these dogs were also reviewed. Materials and Methods Criteria for Patient Selection Medical records of the North Carolina State University (Raleigh, NC) Veterinary Teaching Hospital (VTH) were searched to identify all dogs that were diagnosed with persistent UTIs or reinfections during These dogs were selected on the basis of having at least

2 Recurrent UTIs in Dogs 63 UTIs confirmed by positive aerobic bacterial cultures and antimicrobial susceptibility (C/S) during a 6-month period. 7 A bacterial culture was considered positive if urine samples were obtained by catheterization or cystocentesis and had 10,000 and 1,000 colonyforming units per milliliter, respectively. Dogs for which follow-up information was not available for the entire 6-month period after the initial UTI were excluded from the study. Data Collection Signalment. Sex, age at 1st UTI, month of initial presentation, and breed were recorded. Each dog was assigned to one of the following 7 American Kennel Club defined breed groups 0 : sporting dogs, hounds, working dogs, terriers, toys, nonsporting dogs, or herding dogs. A separate group was assigned for mixed-breed dogs. King Charles Cavalier Spaniels were assigned to the sporting dogs group with the other spaniels. Australian Sheepdogs and Border Collies were assigned to the herding dogs group. The reproductive status was recorded as intact or neutered on the basis of the dog s status at the time of 1st presentation for a UTI. A computer-performed random selection was made of 1st-accession records from dogs presenting to the VTH during the same 10-year period to generate a control population of 400 dogs. Breed, age, sex, and reproductive status of these 400 dogs at the time of 1st presentation to the VTH were obtained by conducting a computer-assisted search of the VTH medical records. Age was rounded to the nearest year for all dogs except for those that were younger than 1 year, whose ages were described as fractions of a year. The signalments of dogs with UTIs were compared to the signalments of dogs in the control population. Presenting Complaint. Information about the presenting complaint, previous UTIs diagnosed by the referring veterinarian, and historic and physical findings consistent with a UTI was extracted from the medical record for the 1st visit from which a positive urine culture was recorded. Historic and physical findings consistent with a UTI included inappropriate urination, dysuria, hematuria, stranguria, pollakiuria, or malodorous urine. Dogs were classified as follows: (1) presenting with a complaint of frequent UTIs; () presenting for signs consistent with a UTI; (3) presenting for another problem but with concurrent signs consistent with a UTI; or (4) presenting for another problem and asymptomatic for a UTI. Urinalyses Results. The following information was recorded from the 1st (n 88) and last urinalyses in the record associated with positive culture results: method of collection, specific gravity, presence or absence of nonhyaline casts, presence or absence of bacteriuria, and degree of pyuria or hematuria. For statistical analysis, hematuria and pyuria were grouped as 0 5 (normal), 5 50, and 50 cells per highpower field. Not all urine samples that were submitted for culture were also submitted for urinalysis. Data from samples collected by voiding were excluded. Bacterial Isolates and Antimicrobial Susceptibility Patterns. Urine samples were cultured aerobically by standard methods: samples were centrifuged, and the pellet was inoculated onto plates containing Columbia agar with 5% sheep blood, phenylethyl alcohol agar, and deoxycholate agar as well as into liquid thioglycolate medium. Before centrifugation, a 100-fold dilution of urine also was mixed with molten agar and poured to estimate bacterial numbers. All media were incubated at 35 C overnight. Bacterial species were identified by an automated system. a Antimicrobial susceptibility patterns for each urine C/S were determined by the VTH Microbiology Laboratory with their automated system for the analysis of most bacterial species and by the Kirby-Bauer technique for the analysis of Streptococci. The susceptibility panel used was an extended panel that varied depending on the year of the culture and the isolate being tested. The number of different isolates and the species of those bacterial isolates from each culture were recorded. Bacterial isolates and susceptibility patterns were recorded for both the 1st culture in each record and separately for all 373 cultures obtained from the 100 dogs. The degree of resistance was categorized into 5 groups as follows: resistant to all antibiotics tested; resistant to all antibiotics except for those that are available for injection only; resistant to all commonly prescribed antibiotics administered PO but susceptible to at least one nd- or 3rd-generation cephalosporin administered PO; resistant to all commonly prescribed antibiotics administered PO but susceptible to at least 1 fluoroquinolone; or susceptible to at least 1 commonly prescribed antibiotic administered PO. Commonly prescribed antibiotics that were administered PO were defined as those typically used in the treatment of UTIs. These included trimethoprim-sulfa, amoxicillin with clavulanic acid, amoxicillin, cephalexin, and tetracycline. Fluoroquinolones and nd- and 3rd-generation cephalosporins were considered secondary antibiotics. The degrees of resistance of isolates from 1st-time infections cultured during and were recorded to allow a comparison of patterns of antibiotic resistance over the course of the study. A comparison was made of only those isolates that came from the 1st UTI of record to minimize the effect of prior antibiotic administration on the degree of resistance, but 9% of the affected dogs came with a prior diagnosis of a UTI and had therefore received some antibiotics before referral. Possible Risk Factors. Medical records were searched for evidence of disorders that have the potential to be associated with the development of a complicated UTI. The disorders were grouped for analysis according to previously defined classes 1 as follows: abnormal micturition (eg, detrusor dysergia, prostatic abscesses, cysts), abnormal anatomic structure (eg, vestibulovaginal stenosis), abnormal mucosal defense barrier (eg, transitional cell carcinoma), altered urine composition (eg, diabetes mellitus), abnormal systemic immunocompetence (eg, hyperadrenocorticism, corticosteroid therapy), or multiple classes present. Whether or not the disorder was corrected also was recorded. Treatment Strategies. Treatment strategies were recorded and grouped as follows: correction of predisposing disorder (eg, surgical removal of uroliths, management of hyperadrenocorticism) with or without antibiotic or ancillary therapy; ancillary therapy (eg, urinary antiseptics, methenamine mandelate) with or without concurrent antibiotic therapy; long-term, low-dose antibiotic therapy; or standard antibiotic therapy alone. Long-term, low-dose antibiotic therapy was defined as the use of a subtherapeutic dosage of an antibiotic (typically one third of the standard dosage) given nightly to prevent reinfection after the clearance of infection had been confirmed by a negative urine culture. When multiple strategies were used throughout the dog s clinical course, the strategy used for the purpose of analysis was recorded in the following order of preference: (1) correction of the predisposing disorder; () ancillary therapy; (3) long-term, low-dose antibiotic therapy; or (4) standard antibiotic therapy alone. To evaluate treatment efficacy, the interval until the next UTI (apparent disease-free interval [ADFI]) was recorded in addition to the number of UTIs within the 6-month period after the initiation of the treatment. Urine cultures were not performed at regular intervals, precluding the calculation of a true disease-free interval. The ADFI was calculated in weeks. If the follow-up urine culture results were compatible with the persistence of the original UTI, the ADFI was recorded as zero. A dog was determined to be disease free if a followup urine culture did not yield bacterial growth. The ADFI of dogs with a negative follow-up culture was determined as the time from the 1st UTI until the next positive culture. In situations in which the followup urine culture was positive but suggestive of reinfection (ie, new organism cultured, original organism no longer present), the dogs were presumed to have been disease free for some time after treatment of the original infection before reinfection. In these dogs, the ADFI was calculated as the period between the 1st and nd positive urine culture. For purposes of analysis, the ADFI was grouped as follows: 0 weeks, 8 weeks, 8 4 weeks, and 4 weeks. Type of Infection. Each infection was classified on the basis of C/S results, when possible, as one of the following: the 1st UTI identified at the VTH, relapse, or persistence of previous infection (ie, same

3 64 Seguin et al Table 1. Signalment (age, sex, and breed grouping) and season of 1st presentation of dogs with reinfections, persistent urinary tract infections, or both and of dogs in the hospital reference population. UTI Population (n 100) Reference Population (n 400) Variable No./% Age (years) (5.) 78 (19.5) 140 (35.0) 161 (40.) Sex Male/intact Male/castrated Female/intact Female/spayed (33.0) 80 (0.0) 57 (14.) 131 (3.7) Breed grouping Mixed breed Sporting breeds Hounds Working breeds Terriers Toy breeds Nonsporting breeds Herding breeds (6.7) 8 (0.5) 8 (7.0) 36 (9.0) 9 (7.) 50 (1.5) 7 (6.7) 41 (10.) Season January March April June July September October December (0.5) 135 (33.8) 105 (6.3) 78 (19.5) UTI, urinary tract infection. organism, same antimicrobial susceptibility pattern); reinfection (new organism or different antimicrobial susceptibility pattern); or superinfection (previous organism still present and additional organism). Statistical Analysis The continuous variable age initially was categorized and coded into previously documented biologically relevant age groups. 1 Sample size restrictions, however, necessitated further grouping of the dogs into 4 age categories: younger than 3 years, 4 6 years, 7 10 years, and older than 10 years. The month of admission was categorized by season: winter (January March), spring (April June), summer (July September), and fall (October December). Weight data were not consistently available for the reference group population and were excluded from analysis. Breed was categorized for analysis as noted above. The age, sex, breed groupings, and season of admission of dogs from the UTI population and of dogs from the VTH population were independently compared by logistic regression. b,1, All variables then were included in a series of multivariate models to adjust for the potential relationship between individual variables. Variables initially were added in a forward fashion to improve model validity or later deleted to improve model precision. Changes in the likelihood ratio statistic and odds ratios and corresponding confidence intervals calculated from the beta coefficient 3 were evaluated to assess the contribution of each variable as it was either added to or deleted from subsequent models. Differences in univariate and multivariate odds ratios and corresponding confidence intervals for variables were used to identify confounding variables. The demographic data from animals with and without UTI data were further stratified for analysis by means of logistic regression to examine the relationship of sporting breeds, sex, and age. Similar analysis of other breed groups was limited by sample size constraints. Additional stratified analysis was conducted by calculating odds ratios and 95% confidence intervals to assess the potential association of selected independent variables such as specific gravity with the presence of reinfections or persistent UTIs. Results One hundred dogs during matched the selection criteria described. Six additional dogs met most of the selection criteria but were excluded from the study because they died within 6 months of their initial positive urine culture, thus not allowing adequate follow-up. From the records of the 100 dogs, 374 UTIs were diagnosed by a positive C/S of urine samples obtained by catheterization (n 41) or cystocentesis (n 333). The number of positive cultures per dog ranged from to 19 (median, 3; 5% percentile [Q1], ; 75% percentile [Q3], 4). The time between positive urine cultures in reinfections after the initiation of treatment ranged from to 174 weeks (median, 0; Q1, 8; Q3, 38). The time between positive cultures in persistent infections ranged from to 77 weeks (median, 5; Q1, 4; Q3, 7). Signalments of the dogs with persistent UTIs, reinfections, or both and of the control dogs were recorded (Table 1). The age of dogs at 1st presentation ranged from 0.3 to 16 years (median, 7 years; Q1, 5; Q3, 10). Dogs in the reference population ranged in age from 1 to 0 years (median, 9 years; Q1, 7; Q3, 1). UTIs were diagnosed in 38 breeds (including mixed breeds). Common breeds (those accounting for 5%) within the UTI population included mixed-breed dogs (15%), Golden Retrievers (8%), Dachshunds (8%), Cocker Spaniels (6%), and Labrador Retriev-

4 Recurrent UTIs in Dogs 65 Table. Potential independent association of signalment variables with risk of reinfections or persistent urinary tract infections and final logistic regression model examining the potential association of signalment variables with risk of reinfections, persistent UTIs, or both. Independent Association Final Multivariate Model Variable Odds Ratio 95% Confidence Interval Odds Ratio 95% Confidence Interval Age (years) Sex Breed group Season Male/intact Male/castrated Female/intact Female/spayed Mixed breed Sporting breeds Hounds Working breeds Terriers Toy breeds Nonsporting breeds Herding breeds January March April June July September October December.38* ** *.35* * ** * 3.9* 3.69* * UTIs, urinary tract infections. * Indicates more likely (P.05) to be in the UTI patient population. ** Indicates less likely (P.05) to be in the UTI patient population. ers (5%). Dogs with UTIs were examined as patients at the VTH at a relatively uniform frequency throughout the year. Logistic regression analysis was used to compare the age, breed group, sex, and season of admission of dogs with UTIs and dogs in the hospital reference population (Table ). Although certain breed groups (sporting dogs, hounds, and nonsporting dogs) appeared to be at an increased risk of UTIs, substantial confounding was observed between these breed groups and other signalment variables. Sporting breeds of dogs with reinfections, persistent UTIs, or both were more likely to be younger than 3 years (OR, 7.9; 95% CI, ) and spayed females (OR, 3.50; 95% CI, ). Sample size restrictions limited an additional stratified analysis of the relationship between breed groupings, age, and sex. No association was found between the season of presentation and the likelihood that individual dogs were in the UTI patient population. In this study, 9% of the dogs presented with a primary complaint of repeat UTIs. Nine percent presented with signs of lower urinary tract disease but had no previous history of UTIs, whereas 6% presented for evaluation of problems unrelated to the urinary tract. Slightly more than half of the 100 dogs (54%) were asymptomatic for UTIs at the time of their 1st presentation. Of the dogs with UTIs presenting for a problem unrelated to the urinary tract, only 8 of 6 (1.9%) had clinical signs of a UTI at the time of diagnosis. Urinalysis results were available for 9 of the 374 culture-positive samples. When available, the 1st (n 88) and last (n 80) urinalyses that were associated with positive cultures in each record were analyzed separately for specific gravity and sediment characteristics. The majority (77 of 88, 88.6% for the 1st urinalysis and 77 of 80, 96.3% for the last urinalysis) of these samples were collected by cystocentesis. The remaining 11.4% (1st urinalysis) or 3.8% (last urinalysis) were collected by catheterization. The specific gravity in 1st-occurrence urine samples ranged from 5 to 1.053, but 75% had a specific gravity (mean, 1.0; median, 1.00; Q1, 1.013; Q3, 1.09). Dogs that had a specific gravity 1.05 were less likely to have E coli infection (OR, 0.18; 95% CI, ). Dogs with specific gravities 8 were no more likely than dogs with specific gravities 8 to have 3 recorded UTIs (OR, 1.7; 95% CI, ). The sediment characteristics were highly variable (Table 3). Of all urinalyses (n 9), 1 (4.1%) had unremarkable sediment examinations (defined as no bacteria, no casts, 0 5 red blood cell counts, and 0 5 white blood cell counts per high-power field) despite a positive urine culture. An additional 5 urinalyses (1.7%) had only microscopic hematuria, and 7 (9.%) had only bacteriuria (1 ). E coli was the most common bacterial isolate (Table 4). Of the 373 bacterial isolates, 63 (70.5%) were susceptible to at least 1 commonly prescribed antibiotic administered PO, and 89 (18.5%) were resistant to all commonly prescribed antibiotics administered PO but were susceptible to a secondary antibiotic (41 or 11% susceptible to a nd- or

5 66 Seguin et al Table 3. Urine sediment characteristics of the urinalysis associated with the 1st (n 88) and last (n 80) positive culture in each record. Urine Sediment Hematuria (RBC/hpf) Pyuria (WBC/hpf) Bacteriuria (0 4 ) Degree st UTI 40 (45.5) 9 (33.0) 19 (1.6) (5.0) 9 (33.0) 37 (4.0) 13 (14.8) 15 (17.0) 18 (0.5) 6 (9.5) 16 (18.) Last UTI 5 (65.0) 16 (0.0) 1 (15.0) 3 (8.8) 38 (47.5) 19 (3.8) 1 (15.0) 10 (1.5) 13 (16.3) 30 (37.5) 15 (18.8) Presence of nonhyaline casts 6 (6.8) 3 (3.8) hpf, high-power field; RBC, red blood cell count; UTI, urinary tract infection; WBC, white blood cell count. Table 4. Distribution of bacterial isolates (n 441) from 373 positive urine cultures in dogs with reinfections, persistent UTS, or both at the North Carolina VTH during Pathogen Escherichia coli Enterococcus spp. Proteus spp. Klebsiella spp. Staphylococcus spp. Pseudomonas spp. Streptococcus spp. Enterobacter spp. Other Mixed infections No. Isolates (n) VTH, Veterinary Teaching Hospital. % Cultures (n 373) % Isolates (n 441) rd-generation cephalosporin PO; 48 or 1.9% susceptible to a fluoroquinolone; and 69 or 18.5% susceptible to either). Resistance to all antibiotics except for injectable ones was observed in 8 (7.5%) of the isolates, and complete resistance was observed in 15 (4%) of the isolates. Eleven of the 15 cultures that were resistant to all antibiotics were mixed infections, and all 15 included an Enterococcus spp. The antimicrobial susceptibility patterns of the different bacterial species isolated from single-isolate UTIs in this study were based on the C/S results. Isolates that were relatively susceptible to commonly prescribed antibiotics administered PO included Proteus mirabilis (100% susceptible to a commonly prescribed antibiotic administered PO), Staphylococcus spp. (94% susceptible), and Klebsiella pneumoniae (84% susceptible). Moderately susceptible isolates included E coli (7.5% susceptible) and Streptococcus spp. (including Enterococcus spp., 70% susceptible). Pseudomonas spp. were notably resistant, with only 1 of 11 isolates being susceptible to a commonly prescribed antibiotic administered PO. Forty-five percent of the mixed infections contained at least 1 isolate that was resistant to all commonly prescribed antibiotics administered PO. Bacterial isolates obtained at the beginning of the study appeared to have a lesser degree of resistance than bacteria cultured near the end of the study. During , 1 urine cultures were submitted from dogs during their 1st recorded UTI. Of these, 0 of 1 (95.%) were susceptible to at least 1 commonly prescribed antibiotic administered PO. The remaining isolate was susceptible to a secondary antibiotic administered PO (a nd- or 3rd-generation cephalosporin). By contrast, in , only 59.% of 1st-infection isolates (n 7) were susceptible to a commonly prescribed antibiotic administered PO. Six isolates (.%) were susceptible to a secondary antibiotic administered PO but not to a commonly prescribed antibiotic. Five of the remaining isolates were resistant to all of the antibiotics administered PO, with of 7 (7.4%) resistant to all exclusively injectable antibiotics tested as well. Insufficient numbers of isolates were available for statistical comparison. A search of the medical records identified 10 instances of disorders that could have contributed to the complicated UTIs in 71 of the 100 dogs (Table 5). Of 10 instances, there were 18 different disorders. Of the 71 dogs, 1 dogs had multiple disorders that belonged to (n 11) or 3 (n 1) different classes. Fifteen dogs had multiple disorders that did not cross class boundaries. For example, 1 dog was on immunosuppressive corticosteroid therapy for systemic lupus erythematosus (both were classified under impaired immunocompetence). 4,5 No association was found between the disorder class and the number of positive cultures in a 6-month period, an ADFI 8 weeks (Table 6), or the presence of pyuria. Dogs with abnormal anatomy, abnormal mucosal barriers, or multiple classes of predisposing disorders were at an increased risk of having positive cultures in their records (Table 7). An association also was noted between abnormal micturition and the degree of antimicrobial susceptibility. Dogs with abnormal micturition were less likely to have isolates cultured from their urine that were susceptible to at least 1 commonly prescribed antibiotic administered PO (OR, 0.14; 95% CI, ). A correction of the identified disorders was performed in 5% of the 100 dogs (35.% of the 71 dogs in which a suspected predisposing disorder had been identified). Of these, 1 dog also was treated by low-dose, long-term antibiotic therapy, and dogs were given methenamine mandelate. All other dogs in which an underlying disorder was corrected were given adjunctive standard-dose, short-term antibiotic therapy. In some dogs, a correction of the disorder was not attempted because it would have been detrimental to the patient (eg, cessation of immunosuppressive therapy for the treatment of lymphosarcoma) or was otherwise not feasible (eg, surgical correction of transitional cell carcinoma of the bladder trigone). Methenamine mandelate was used in 4 other dogs for which a correction of the predisposing disorder was not attempted. In all dogs in which methenamine mandelate was used, the bacterial isolate (1) was resistant to all antibiotics tested, () was resis-

6 Recurrent UTIs in Dogs 67 Table 5. Distribution of disorders (n 10) identified in 71 dogs a with reinfections, persistent UTIs, or both that may have contributed to the development of UTIs, separated by class. Class of Disorder Abnormal micturition (n 15 dogs) Anatomic defects (n 4 dogs) Alteration of urothelium (n 0 dogs) Altered urine composition (n 7 dogs) Impaired immunity (n 36 dogs) Specific Condition Paraparesis Detrusor dysnergia Prostatic cysts or abscesses Urethrostomy Urethral avulsion/rupture Ectopic ureters Incontinence, undetermined cause Pelvic bladder with incontinence Vestibulovaginal stenosis Hooded vulva Transitional cell carcinoma Urolithiasis Hypoadrenocorticism Diabetes mellitus Corticosteroid therapy Chemotherapy administration Systemic lupus erythmatosis Hyperadrenocorticism No. Dogs Affected UTIs, urinary tract infections. a Twenty-seven dogs had multiple disorders. Potentially predisposing disorders were not identified in 9 dogs. tant to all antibiotics administered PO, or (3) had previously not responded to antibiotics to which it was susceptible in vitro. Low-dose, long-term antibiotic therapy was used in 1 dogs in which a correction of the predisposing disorder was not attempted. The remainder (59%) of the dogs were given standard-dose, short-term antibiotic therapy for all UTIs in their record. The length of antibiotic treatment ranged from 0.5 to 0 weeks (median, 3; Q1, ; Q3, 4). The median, Q1, and Q3 were identical for the 1st positive culture in the record (n 100) compared to that of all positive cultures in the record (n 336) for which that information was available, indicating that antibiotics were not administered for a shorter duration of time in the management of the 1st UTI than for the later UTIs. On the basis of urine C/S results from posttreatment urine samples, 40 (4.1%) of the UTIs were not eliminated (ADFI 0 weeks) after the initiation of the final treatment strategy. The ADFI was 8 weeks in 56 (58.9%) dogs, Table 6. Association a of apparent disease-free interval (ADFI) in dogs with reinfections, persistent UTIs, or both with class of potentially predisposing disorder or treatment strategy. No. % Corrected ADFI ( 8 weeks) ADFI (8 4 weeks) ADFI ( 4 weeks) Class of disorder: None identified Abnormal micturition Anatomic effects Altered urothelium Altered urine composition Impaired immunity Mixed classes 5 b b 19 1 N/A c (64.0) 7 (63.6) 6 (50.0) 9 (69.) 1 (33.3) 11 (57.9) 6 (50.0) 6 (4.0) 3 (7.3) (16.7) 1 (7.7) 0 (0.0) (10.5) 1 (8.3) 3 (1.0) 1 (9.1) 4 (33.3) 3 (3.1) (66.6) 6 (31.6) 5 (41.7) Treatment strategy: Correction of disorder Methenamine mandelate Low-dose, long-term antibiotics Standard antibiotic therapy only (50.0) 3 (50.0) 1 (7.7) 41 (74.5) (8.3) 0 (0.0) (15.4) 11 (0.0) 10 (41.7) 3 (50.0) 10 (77.0) 3 (5.4) ADFI, apparent disease-free interval; UTIs, urinary tract infections; N/A, not applicable. a No statistical association was found between class and a disease-free interval of 8 weeks (mixed classes and treatment strategy were not evaluated). b Unable to determine the ADFI in 4 dogs in the none identified group and 1 dog in the altered urine composition group. c % Corrected would be 50% if all 4 dogs were included.

7 68 Seguin et al Table 7. Association of class of predisposing disorder with the risk of having UTIs in the record. Class of Disorder None identified Abnormal micturition Anatomic effects Alteration of urothelium Altered urine composition Impaired immunity Mixed classes UTIs, urinary tract infections. Odds Ratio % Confidence Interval between 8 and 4 weeks in 15 (15.8%) dogs, and 4 weeks in 4 (5.%) dogs. The ADFI also was calculated for dogs in each treatment group (Table 6). The low number of patients per treatment group precluded a statistical analysis of these results. An examination of the C/S of the nd positive culture in each record was performed to determine the type of infection. Of these 100 cultures, 4% appeared to represent the persistence or relapse of the previously cultured infection (same bacterial species and identical antimicrobial susceptibility pattern), 50% represented reinfection with a different bacterial isolate, and % represented superinfection. In 6 dogs, the type of infection was unclear. The bacterial isolate in 1 dog could not be determined because the record had been damaged. In the other 5 dogs, a nd culture was not performed for a relatively long period after the treatment of the initial infection. The C/S results in that dog showed the same bacterial species with a susceptibility pattern that was similar, except for resistance to the original antibiotics used. Discussion The mean and median age (7 years) of dogs with persistent UTIs and reinfections in our study was similar to that found in a recent retrospective study (7.7 years) 6 and was slightly less than the hospital reference population mean of 9 years. When evaluated by age groupings, dogs with UTIs were more likely to be younger than 3 years than dogs in the reference population. This finding may have been influenced by the relatively high proportion of dogs with anatomic disorders identified in this study (Table 5). Dogs with UTIs were less likely than dogs in the reference population to be older than 10 years. This observation is counterintuitive, because as a dog s age increases, the likelihood of medical disorders that may predispose the dog to a UTI (eg, diabetes mellitus, hyperadrenocorticism) also increases. However, it may be that dogs developing these medical disorders are less likely to survive to an older age. In this study, UTI patients were more likely to be spayed females than were dogs in the reference population. Studies of uncomplicated UTIs have provided additional evidence that female dogs are at an increased risk for UTIs. 1,6,7 Some studies have found an increased risk of uncomplicated UTIs in spayed females compared to intact female dogs, 1,10 whereas others have not confirmed this finding. 8 In a recent study of persistent UTIs and reinfections, sex differences were noted for certain breeds, but no specific association between sex and UTIs was observed. 6 However, the reproductive status of the dogs was not evaluated, and the dogs of that study were compared with dogs presenting with 1 episode of an acute UTI rather than with the general population. Between 5 and 35% of women aged 0 40 years are diagnosed with UTIs. 8 A study of sexually active young women found an incidence rate of per person-year. 9 By comparison, one study found that men between the ages of 1 and 50 have a UTI incidence rate of per person-year. 30 This observation may be due to sex differences in the anatomy of the urethra, but hormonal influences on urine composition and antibacterial properties of prostatic secretions also may play a role. 6,31 Anatomic considerations, especially urethral length, also may contribute to the more common occurrence of UTIs in female dogs than in males. Only 100 dogs representing 38 breeds were examined in this study, and the small number of dogs of specific breeds precluded a more comprehensive assessment of breed association with recurrent UTIs. An analysis of breed groupings, however, determined that sporting breed group dogs (which includes retrievers and spaniels), hounds, and nonsporting breeds were more likely to be present in the UTI patient population than in the reference population. However, substantial confounding between the breed grouping categories, age, and sex occurred. A comparison of the C/S results with those obtained previously allowed an approximate classification of infections as persistent, reinfection, or superinfection. Dogs were almost as likely to have failed to clear the infection despite therapy (4%) as to have acquired an infection with a new organism (50%). This finding illustrates the importance of repeated cultures after antibiotic therapy to confirm that the infection has been eradicated in dogs suspected of having a complicated UTI. This classification was limited because not all dogs had follow-up urine cultures submitted during or immediately after antibiotic therapy. Therefore, some of the UTIs that were classified as persistent (ie, if the same bacterial species with the same susceptibility pattern on subsequent urine cultures was present) actually may have been reinfected by the same bacterial strain. The relative frequency of different bacteria in the cultures of this population of dogs was similar to that reported in studies of reinfections, persistent UTIs, or both 6,10 1 and in large studies of urine cultures that did not distinguish between solitary and repeat infections Overall, the relative frequencies of bacterial isolates have not differed substantially in reports exclusively evaluating complicated, persistent UTIs, reinfections, or both compared with studies including dogs with all forms of UTIs. A study of complicated UTIs associated with urolithiasis is an exception. 10 In this study, cultures from both urine and uroliths were included in the analysis. The high frequency of Staphylococcus spp. isolates observed may have reflected the cultures of uroliths. This observation is consistent with the findings of another study, which has shown an increased frequency of Staphylococcus spp. in cultures of struvite uroliths. 35 Bacteria isolated from the dogs of our study appeared to be less susceptible to commonly prescribed antibiotics ad-

8 Recurrent UTIs in Dogs 69 ministered PO than did isolates in previous reports of antimicrobial susceptibility patterns from UTIs in dogs. 36,37 This difference was particularly notable for Pseudomonas, Streptococcus, and multiple-isolate infections. Pseudomonas spp. have previously been reported to be 80% susceptible to tetracyclines. Likewise, Streptococcus spp. have been reported to be 100% susceptible to penicillin and ampicillin. 34,36 However, the susceptibility results reported in our study were based on the susceptibility of isolates to serum concentrations of antibiotics rather than the urinary concentrations of antibiotics, which for some antibiotics can be much higher than those in serum. 36 The use of serum concentrations may therefore result in an overestimation of resistant UTIs. The increased resistance seen in the dogs of this report may also reflect a change in the antibiotic resistance patterns of all infections over time. The apparent increased resistance noted in the initial isolates from later in the study compared to those from the 1st years is suggestive of a trend toward increased resistance over time. Further work must be performed to examine antimicrobial susceptibility over time and in relation to reinfections and persistent UTIs. Urinalysis characteristics of the population of dogs of our study were highly variable. Clinically relevant ( 5 cells per high-power field) hematuria and pyuria were absent in 53 and 8% of the urinalyses, respectively. Bacteriuria, identified in 85% of the urine samples, was the most consistent finding. In normal dogs, bacteriuria in the absence of pyuria is often considered suggestive of contamination of the urine sample rather than an actual infection. In the present study, bacteriuria in the absence of any other sediment changes was a fairly common finding in urinalyses from these culture-positive dogs. Microscopic hematuria in the absence of other sediment changes (seen in 1.7% of the dogs of our study) often is attributed to cystocentesis-induced trauma. Urine sediment examinations were unremarkable in 4.% of the urinalyses. Overall, a total of 15% of the urinalyses had either a negative sediment or only minor abnormalities (eg, 1 bacteriuria or microscopic hematuria), which could have reduced the index of suspicion for UTIs in these dogs. Urine samples collected from dogs infected with E coli were significantly more likely to have a specific gravity 1.05 than those collected from dogs with UTIs caused by other bacterial isolates. This finding may reflect endotoxin-mediated effects on urinary concentrating ability. Alternatively, it may reflect the increased susceptibility of E coli to the antibacterial or osmotic properties of concentrated urine. Dilute urine in these dogs also may have reflected their underlying disease processes. For example, hyperadrenocorticism or corticosteroid therapy may lead to more dilute urine and a decreased inflammatory response to infection. Dogs with disorders known to predispose them to UTIs should periodically have C/S testing of their urine samples collected by cystocentesis, regardless of their urinalysis findings. A wide range of disorders suspected to negatively affect host defenses against UTIs were identified in the dogs of this study. Many of the dogs with persistent UTIs or reinfections were paraplegic or were receiving chemotherapy, corticosteroids, or both medications. These findings probably reflect the referral nature of our hospital. Anatomic defects and urolithiasis were commonly identified disorders in the UTI population. Potentially predisposing disorders were not identified in 9% of the dogs with persistent UTIs or reinfections, but a systematic search for predisposing causes was not conducted in every dog. Dogs with abnormal anatomic structures, abnormal mucosal defense barriers, or multiple classes of disorders were more likely to develop 3 UTIs than were other dogs in the UTI population. However, no association was found between the class of predisposing disorders and the ADFI or the number of UTIs in a 6-month period. The increase in the total number of UTIs in the record may be a reflection of the more chronic and long-term nature of these particular underlying disorders rather than a reflection of a predisposing effect compared to other classes of disorders. Dogs that have been on chemotherapy, for example, may have had compromised systemic immunity and altered urine composition (if chemotherapy included steroid therapy), leading to reinfections and persistent UTIs for only a short time before they died as a consequence of their primary disease. Their prognosis would limit their chances of developing a large number of UTIs. Dogs with an abnormal anatomy, by contrast, may achieve a normal life span even without a correction of their disorder. A controlled study is needed to determine the relative predisposing effects of the different disorders. Treatment strategies for persistent UTIs and reinfections include removal or appropriate management of the predisposing disorder when identified; administration of antibiotics for a longer duration (3 6 weeks) than is recommended for a simple UTI; long-term, low-dose antibiotic therapy; and ancillary treatments such as methenamine mandelate or urinary acidifiers. The relative success of these strategies in dogs has not, to our knowledge, been reported previously. Even when there are reinfections or persistent infections, conventional antibiotic therapy is the most frequently chosen treatment option. Ideally, the correction of predisposing disorders should be the mainstay of management of dogs with complicated UTIs. However, a correction of the predisposing problem often was not possible in the dogs of this report. Attempts to correct identifiable disorders were made in 35.0% of the dogs. A subsequent ADFI 4 weeks was seen in 41.7% of these dogs compared to 5.4% of the dogs treated only with conventional antibiotic therapy. This finding suggests that the disorders that were identified and corrected were contributing to the development of reinfections, persistent infections, or both in many of the dogs and that the correction of such disorders was warranted. Only 8.3% of the dogs in which a predisposing disorder was corrected developed reinfection in the 8- to 4-week period after correction (compared to 0% in that same period after antibiotic treatment alone). However, 50% of these dogs had a nd positive urine culture within the 1st 8 weeks of an attempted correction of the underlying disorder. Although less than in the standard antibiotic group (74.5%), this frequency of positive cultures is more than was expected and may reflect sample collection during the recovery period or failure to adequately correct all underlying disorders. Of all the treatment strategies analyzed, low-dose, long-term antibiotic therapy seemed to

9 630 Seguin et al be the most effective. Fewer than 8% of the dogs receiving this treatment had a positive culture within the 1st 8 weeks after the initiation of therapy. Although these dogs had frequent urine cultures during low-dose antibiotic therapy, 77% did not have a positive urine culture within the 1st 4 weeks of treatment. Most of the dogs of this study that presented with reinfections or persistent infections were initially presented for evaluation of another problem. Of these dogs, 87% had no historic or physical evidence of a UTI. In addition, the urine sediment results were not always suggestive of infection. Up to 15% of the urine sediments could easily have been misinterpreted as normal had a C/S not been performed. Microbial cultures from these dogs identified a high rate of resistance, with only 70.5% of the isolates susceptible to a commonly prescribed antibiotic administered PO. This finding emphasizes the importance of performing a urine culture on dogs for which a high degree of suspicion exists in the presence of a potentially predisposing disorder. Despite an appropriate antibiotic treatment that was based on the results of the C/S, the dogs of this study had a high rate of persistent infections. A large percentage of the UTI patient population with repeat infections had at least 1 disorder that may have contributed to the reinfection, the persistence of UTIs, or both. As expected, when anatomic or medical disorders were corrected, the dogs appeared to do better overall, with fewer infections noted. However, only 100 patients were available for study, and a great deal of variability was observed in the management of the dogs. When a correction of potentially contributing disorders is not possible, lowdose, long-term antibiotic therapy may represent a more useful strategy than short-term antibiotic therapy. Prospective, controlled studies are needed to determine the effectiveness of low-dose, long-term antibiotics as a strategy for the treatment of reinfections and persistent UTIs. Footnotes a VITEK automated system, biomerieux, St Louis, MO b Egret, Cytel Software Corporation, Cambridge, MA References 1. Osborne CA, Lees GE. Bacterial infections of the canine and feline urinary tract. In: Osborne CA, Finco DR, ed. Canine and Feline Nephrology and Urology. Baltimore, MD: Williams and Wilkins; 1995: Osborne CA. Three steps to effective management of bacterial urinary tract infections: Diagnosis, diagnosis, diagnosis. Compend Cont Educ Small Anim Pract 1995;17: Cox CE, Hinman F. Experiments with induced bacteriuria, vesical emptying and bacterial growth on the mechanism of bladder defense to infection. J Urol 1961;86: Gregory JG, Wein AJ, Sansone TC, Murphy JJ. Bladder resistance to infection. J Urol 1971;106:0. 5. Polzin DJ. Management of recurrent bacterial urinary tract infections. Compend Cont Educ Small Anim Pract 1994;16: Norris CR, Williams BJ, Ling GV, et al. Recurrent and persistent urinary tract infections in dogs: 383 cases ( ). J Am Anim Hosp Assoc 000;36: Schaeffer AJ. Infections of the urinary tract. In: Walsh Campbell s Urology, 7th ed. Philadelphia, PA: WB Saunders; 1998: Freshman JL, Reif JS, Allen TA, Jones RL. Risk factors associated with urinary tract infection in female dogs. Prev Vet Med 1989; 7: Kogika MM, Fortunato VAB, Mamizuka EM, et al. Etiologic study of urinary tract infection in dogs. Braz J Vet Res Anim Sci 1995; 3: Ling GV, Franti CE, Johnson DL, Ruby AL. Urolithiasis in dogs III: Prevalence of urinary tract infection and interrelations of infection, age, sex, and mineral composition. Am J Vet Res 1998;59: Forrester SD, Troy GC, Dalton MN, et al. Retrospective evaluation of urinary tract infection in 4 dogs with hyperadrenocorticism or diabetes mellitus or both. J Vet Intern Med 1999;13: Ihrke PJ, Norton AL, Ling GV, Stannard AA. Urinary tract infection associated with long-term corticosteroid administration in dogs with chronic skin diseases. J Am Vet Med Assoc 1985;186: Biertuempfel PH, Ling GV, Ling GA. Urinary tract infection resulting from catheterization in healthy adult dogs. J Am Vet Med Assoc 1981;178: Barsanti JA, Blue J, Edmunds J. Urinary tract infection due to indwelling urinary bladder catheters in dogs and cats. J Am Vet Med Assoc 1985;187: Smith JD, Stone EA, Gilson SD. Placement of a permanent cystotomy catheter to relieve urine outflow obstruction in dogs with transitional cell carcinoma. J Am Vet Med Assoc 1995;06: Adams WM, DiBartola SP. Radiographic and clinical features of pelvic bladder in the dog. J Am Vet Med Assoc 1983;18: Kyles AE, Vaden S, Hardie EM, Stone EA. Vestibulovaginal stenosis in dogs: 18 cases ( ). J Am Vet Med Assoc 1996; 09: Aldrich J, Ling GV, Ruby AL, et al. Silica-containing urinary calculi in dogs ( ). J Vet Intern Med 1997;11: Batmuzi EK, Kristensen F. Urinary tract infection: The role of canine transmissible venereal tumour. J Small Anim Pract 1996;37: American Kennel Club Staff. The Complete Dog Book: The Official Publication of the American Kennel Club, 19th ed. New York, NY: Howell Book House; Vaden SL, Levine J, Breitschwerdt EB. A retrospective casecontrol of acute renal failure in 99 dogs. J Vet Intern Med 1997;11: Hosmer DW, Lerneshow S. Applied Logistic Regression. Hoboken, NJ: J Wiley; Kleinbaum DG. Logistic Regression: A Self-Learning Text. New York, NY: Springer-Verlag; 1994: Day MJ. Inheritance of serum autoantibody, reduced serum IgA and autoimmune disease in a canine breeding colony. Vet Immunol Immunopathol 1996;53: Chabanne L, Fournel C, Cauz C, et al. Abnormalities of lymphocyte subsets in canine systemic lupus erythematosus. Autoimmunity 1995;: Osborne CA, Klausner JS, Lees GE. Urinary tract infections: Normal and abnormal host defense mechanisms. Vet Clin North Am 1979;9: Kivistö AK, Vasenius H, Sandholm M. Canine bacteriuria. Small Anim Pract 1977;18: Kunin CM. The concepts of significant bacteriuria and asymptomatic bacteriuria, clinical syndromes and the epidemiology of urinary tract infections. In: Detection, Prevention, and Management of Urinary Tract Infections, 4th ed. Philadelphia, PA: Lea and Febiger; 1987: Hooton TM, Scholes D, Hughes JP, et al. A prospective study of risk factors for symptomatic urinary tract infection in young women. N Engl J Med 1996;335: