SEARCH

SEARCH BY CITATION

Keywords:

  • Bacterial morphology;
  • quantitative urinary culture;
  • receiver operating characteristic (ROC);
  • sensitivity;
  • specificity;
  • test accuracy

Background: Urinary sediment examination and quantitative urinary culture results are frequently discordant.

Objectives: The aims of this study were to compare accuracy of light microscopic examination of wet-mounted unstained (wet-unstained) and air-dried modified Wright-stained (dry-stained) sedimented preparations of urine with results of quantitative aerobic bacterial culture for detection and characterization of bacteriuria in cats. In addition, the presence of pyuria detected by urinalysis and potential risk factors were assessed.

Methods: A blinded prospective study was conducted on 472 urinary samples collected from 410 cats by cystocentesis. The age and sex of each cat were recorded. Complete urinalyses were performed and included quantification of WBCs. Quantity and morphology of bacteria in each specimen were determined by light microscopic examination of wet-unstained (performed by certified medical technologists) and dry-stained (performed by a veterinary clinical pathologist) sedimented preparations of urine and compared with results of quantitative bacterial cultures.

Results: Of 472 urinary specimens, 29 were positive for bacteriuria by culture and considered true positives and 443 were considered true negatives. Compared with these results, examination of wet-unstained and dry-stained urines had sensitivities of 75.9% and 82.8%, specificities of 56.7% and 98.7%, and test efficiencies of 57.8% and 97.7%, respectively. Positive likelihood ratios were 1.8 and 63.7 and negative likelihood ratios were 0.42 and 0.17 for wet-unstained and dry-stained examinations, respectively. Compared with 29 culture-positive samples, the wet-unstained method had morphologic concordance and misclassification rates of 37.9% and 62.1%, respectively, whereas the dry-stained method had morphologic concordance and misclassification rates of 65.5% and 34.5%, respectively. Only 34% of samples with bacteriuria had pyuria. Frequency of bacteriuria was not significantly different based on age and sex of the cats, but there was a tendency for increased frequency in female cats and in cats >10 years old.

Conclusions: Staining dried urinary sediment with a modified Wright-stain significantly improved sensitivity, specificity, and test efficiency of microscopic detection and classification of bacteriuria compared with the wet-unstained method. Pyuria should not be a criterion for determining the presence or absence of bacteriuria.