Detection of Occult Urinary Tract Infections in Dogs With Diabetes Mellitus
Dogs with diabetes mellitus may develop occult urinary tract infections. In this study, diabetic dogs with negative and positive bacterial urine cultures were compared. Records from 51 dogs with diabetes mellitus were reviewed at the University of Illinois. No difference was identified between the groups in urine specific gravity, pH, glucose, ketones, protein, red blood cells, white blood cells, or epithelial cells. Dogs with occult urinary tract infection did have an increased incidence of bacteriuria, but this was not a consistent finding. Therefore, the urine on all diabetic dogs should be cultured to accurately identify the presence or absence of bacterial urinary tract infections.
Introduction
Urinary tract infections are frequently identified in dogs with diabetes mellitus. An increased level of white blood cells (WBC) in the urine, or pyuria, defined as >3 WBC per high-power field (hpf) identified on urinary sediment in samples collected by cystocentesis, is the most reliable indicator of bacterial urinary tract infections.1 Urinary tract infections with absence of pyuria are considered to be occult. In immunosuppressed animals, including those with diabetes mellitus, the absence of an inflammatory response to infection may result in an increased incidence of urinary tract infections, many of which may be occult.2–4 Although many veterinary textbooks report the occurrence of occult urinary tract infections in diabetic patients, an extensive literature search revealed only one article describing infections without pyuria in diabetic dogs.3 Individual urinalysis parameters in diabetic dogs were not presented in that study.3
The purposes of the current study were to evaluate for the presence of occult bacterial urinary tract infections in dogs with diabetes mellitus and to determine if urine parameters other than WBC numbers might increase clinical suspicion of urinary tract infections in those with an occult urinary tract infection. The study was performed by comparing urinalysis results from diabetic dogs with no pyuria and negative bacterial urine cultures (noninfected, Group 1) to urinalysis results from diabetic dogs with no pyuria and positive bacterial urine cultures (infected, Group 2).
Materials and Methods
Medical records of dogs with a diagnosis of diabetes mellitus presented to the Veterinary Teaching Hospital at the University of Illinois from January 1995 through May 1999 were reviewed. Diagnosis of diabetes mellitus was made based on clinical signs (i.e., polyuria/polydipsia [pu/pd], polyphagia, weight loss), fasting hyperglycemia, and glucosuria. All diabetic dogs, both newly diagnosed and previously diagnosed, which had a urine sample collected by cystocentesis and a complete urinalysis performed (including sediment examination and bacterial culture and sensitivity) were reviewed for inclusion in this study. Since the purpose of the study was the evaluation of occult urinary tract infections, dogs with pyuria were excluded from the study. Any patient receiving antibiotic medication within 1 week prior to admission was also excluded.
Signalment, history, physical examination findings, concurrent diseases, and results of urinalysis and quantitative bacterial urine culture were extracted from the patient record. Signs consistent with urinary tract disease, including pu/pd, stranguria, and dysuria, were noted.
Urinalyses were performed in a consistent manner for all samples. Urine specific gravity (USG) was determined by refractometer; urine pH, protein, glucose, and ketones were measured using a urine dipstick;a urine sediment was examined microscopically at 40× magnification to evaluate for the presence of red blood cells (RBC), WBC, bacteria, and epithelial cells. Crystals and casts were identified in urine sediments in only four dogs and therefore were not compared between groups in the study. Laboratory technicians, who had been trained to read and quantify urine sediment in a specified manner, read urine sediment.
Quantitative urine cultures were performed by inoculating 0.01 mL of urine onto blood agar and MacConkey’s agar plate. Plates were incubated at 25°C, and a colony count was determined at 24 and 48 hours. A bacterial count of >1 × 103 colony-forming units/mL of urine was considered indicative of bacterial infection.1
Statistical Analysis
Normalcy of the data was evaluated with the proc univariate procedure in SAS.b The data was not normally distributed, and transformation of the data was not successful. Therefore, the Wilcoxon’s one-way nonparametric test was used to analyze the data. A significance level of P<0.05 was employed. The group variables of sex and age and the urinalysis variables of USG, pH, protein, glucose, ketones, RBC count, WBC count, epithelial cell count, and bacterial count were analyzed.
Results
Fifty-one diabetic dogs were included in the study. Of these, 39 (76%) had negative urine cultures (Group 1), and 12 (24%) dogs had positive urine cultures (Group 2). Forty-two (82%) of the 51 dogs were female (three intact, 39 spayed), and nine were male (four intact, five castrated). A greater percentage (11/12; 92%) of females were present in the positive culture group (Group 2). The study included a mix of newly diagnosed and previously diagnosed diabetics. The infected group had two (17%) new diabetic dogs, and the noninfected group had 12 (31%) newly diagnosed diabetics.
Fifteen different breeds were represented in the study. Breeds with multiple animals in the study included miniature schnauzers (n=15), miniature poodles (n=5), Labrador retrievers (n=4), dachshunds (n=3), bichon frise (n=2), and Maltese (n=2).
No significant difference in age was noted between the infected and noninfected groups (P=0.19). The median age for the noninfected dogs was 9.0 years, with a range of 5 to 14 years. Infected dogs had a median age of 7.5 years (range, 3 to 12 years).
Polyuria and polydipsia were the only clinical signs related to the urinary tract present in the dogs of this study. Polyuria/polydipsia was a presenting complaint in 17 (44%) of 39 dogs with negative cultures and in four (33%) of the 12 dogs with positive cultures.
Dogs with hyperadrenocorticism are also known to be predisposed to occult urinary tract infections.13 Seven of the 51 dogs had concurrent hyperadrenocorticism, of which six were being treated with mitotane. Only one of the diabetic dogs with concurrent hyperadrenocortism had an occult infection. In this dog, the hyperadrenocorticism was considered controlled (i.e., adrenocorticotropic stimulation test results were normal) with mitotane therapy.
In the infected dogs (Group 2), the most common bacterial isolate was Escherichia coli (E. coli), identified in five of the 12 dogs with occult urinary tract infections. Other bacterial isolates included Streptococcus alpha in two dogs and Streptococcus delta. Lactobacillus spp., Enterobacter spp., Klebsiella spp., and Pseudomonas spp. isolated from one dog each. Two of the 12 dogs had mixed infections: one with two strains of E. coli and one with E. coli and Klebsiella spp.
When comparing urinalysis results between the two groups of dogs, there was no significant difference in USG or in urine dipstick readings of pH, glucose, protein, or ketones [see Table] between Group 1 (culture-negative) and Group 2 (culture-positive). The median urine pH and glucose in the culture-negative group was 5 and 500 mg/dL, respectively, and 6.5 and 750 mg/dL, respectively, in the culture-positive group. The median for protein (1+) and for ketones (0) did not differ between the groups. No statistical difference was noted in the number of observed cells or the median values for WBC (1/hpf), RBC (1/hpf), and epithelial cells (0/hpf) on urine sediment examination for the culture-negative Group 1 dogs versus the culture-positive Group 2 dogs. However, a significant difference between Group 1 and Group 2 was noted in the number of bacteria per hpf (P<0.02). Group 2 had a greater prevalence of bacteriuria, with 33% (4/12) having bacteriuria, while Group 1 had an 8% (3/39) incidence of bacteriuria.
Discussion
Occult urinary tract infections occurred in 12 (24%) of the 51 dogs in this study. The signalment of the dogs in this study was similar to that previously reported for diabetic dogs.5 Forty-two (82%) of the dogs were female. Eleven (92%) of the 12 dogs with occult infections were female. In humans with diabetes mellitus, an increased incidence of urinary tract infection is noted in women.6 A similar increased prevalence of infection in female canine diabetics may be suggested by results of the present study, but a larger study population is needed to fully evaluate this association.
The median age for the noninfected dogs was 9 years, and for infected dogs it was 7.5 years, with no statistically significant difference between the groups. Median age is similar to that reported for dogs with diabetes mellitus, which has a peak incidence between 7 and 9 years of age.5
Polyuria/polydipsia was the only clinical sign related to the urinary tract exhibited by dogs in this study. The incidence of pu/pd was similar between dogs with negative and positive urine cultures. The pu/pd was most likely a sign of uncontrolled diabetes mellitus and obligatory glucosuria and was not necessarily related to urinary tract infections.
Bacteriuria was the only urinalysis variable that was significantly different between Groups 1 and 2 and could therefore facilitate differentiation of infected versus noninfected diabetic dogs without pyuria. When comparing dogs in Group 1 to Group 2, there was no statistical difference in USG, chemistries, or number of RBC, WBC, or epithelial cells on urine sedimentation.
Bacterial isolates from the urine cultures were similar to those reported in nondiabetic dogs, with 75% of isolates in this study being gram negative.1 Escherichia coli was the most common isolate in this study, infecting five (42%) of the 12 dogs. Forrester also found E. coli to be the most prevalent organism, causing 69% of urinary tract infections in dogs with diabetes mellitus, hyperadrenocorticism, or both.3 Gram-positive organisms were isolated from 25% of dogs in the present study. Two dogs had mixed infections: one with two strains of E. coli and one with E. coli and Klebsiella sp. In Forrester’s study, 17% of dogs had one or more organisms isolated from the urine in dogs with diabetes mellitus, hyperadrenocorticism, or both.3
Historically, glucosuria in patients with diabetes mellitus was considered an important factor predisposing to the development of urinary tract infection. However, Patterson and Andriole found no difference in growth of bacteria cultured in human urine with and without added glucose, suggesting that the presence of glucose in urine may not be the primary determinant of predisposition to urinary tract infections in diabetics.6 However, high levels of urinary glucose hinder leukocyte phagocytic function predisposing to urinary tract infection.7 The presence of glycosuria was not different in the dogs with and without infections in this study.
In the present study, 12 (24%) of 51 diabetic dogs examined had bacterial urinary tract infection with no evidence of pyuria. Lack of pyuria may result from altered function of granulocytes in diabetics. Multiple leukocyte dysfunctions in adherence, chemotaxis, phagocytosis, and bactericidal activity have been found in human diabetics.89 In poorly controlled diabetic dogs, systemic neutrophil adherence is significantly decreased, though well-regulated diabetic dogs have no alteration in adherence.410 Intracellular bactericidal activity of neutrophils is diminished in diabetic mice with urinary tract infections.11 Altered function of leukocytes may account for the increased incidence of urinary tract infection, including occult infections, in patients with diabetes mellitus.
Conclusion
Occult urinary tract infections (without pyuria) occur relatively frequently in dogs with diabetes mellitus. Therefore, pyuria is not a reliable indication of urinary tract infection in this population. The prevalence of bacteriuria was increased in diabetic dogs with occult infections; however, bacteriuria was not a consistent finding and hence may be unreliable in predicting presence or absence of infection. Consequently, urine cultures should be performed routinely in all diabetic dogs regardless of the presence or absence of clinical signs or urinalysis results suggestive of urinary tract infection.
Chem Strip 9; Boehringer Mannheim Corporation, Indianapolis, IN
SAS user’s guide version 8; SAS Institute, Inc., Cary, NC
