Evaluation of different sampling methods and criteria for diagnosing canine urinary tract infection by quantitative bacterial culture.

The use of voided urine specimens for bacteriological culture in dogs is discouraged because contamination from external genitalia could lead to misinterpretation of laboratory results. Quantitative culturing and defining significant bacteriuria could increase the usefulness of voided specimens. However, limited evidence exists for the cut-offs currently recommended. The aim of this study was to evaluate the accuracy of current veterinary cut-off values for significant bacteriuria in voided canine urine. A secondary aim was to investigate if accuracy improved when applying qualitative criteria used in humans. Paired urine specimens were collected by both cystocentesis and voiding, and quantitative bacteriological cultures were performed within the same day. Cystocentesis was used as the reference standard with a cut-off for significant bacteriuria of ≥1000 colony forming units (CFU)/mL. Voided specimens were compared to cystocentesis using: (1) the veterinary cut-off of ≥100,000 CFU/mL; and (2) various cut-offs depending on qualitative criteria (sex, clinical signs and complicating factors), adapted from human guidelines. Ninety-four dogs with suspected urinary tract infection (UTI) were included for analysis. The veterinary cut-off yielded an accuracy of 94% with a sensitivity and specificity of 94% (95% confidence intervals [CI] 0.81, 0.99) and 94% (95% CI 0.86, 0.98), respectively. Applying the human guidelines did not improve overall accuracy (89%), and yielded a sensitivity and specificity of 97% (95% CI 0.86, 1.00) and 86% (95% CI 0.77, 0.92), respectively. The veterinary cut-off value of ≥100,000 CFU/mL for voided urine is appropriate for determining significant bacteriuria in the majority of dogs with suspected UTI if specimens are refrigerated and cultured on the day of collection.