Accuracy of urine dipstick tests and urine glucose-to-creatinine ratios for assessment of glucosuria in dogs and cats.

Abstract

To assess the accuracy of automated readings of urine dipstick results for assessment of glucosuria in dogs and cats, compare visual versus automated readings of urine glucose concentration, and determine the utility of the urine glucose-to-creatinine ratio (UGCR) for quantification of glucosuria. 310 canine and 279 feline urine samples. Glucose concentration was estimated in 271 canine and 254 feline urine samples by visual assessment of urine dipstick results and with an automated dipstick reader. Absolute urine glucose and creatinine concentrations were measured in 39 canine and 25 feline urine samples by colorimetric assay with a clinical chemistry analyzer (reference standard for detection of glucosuria), and UGCRs were determined. Automated assessment of the urine dipsticks yielded accurate results for 163 (60.1%) canine urine samples and 234 (92.1%) feline urine samples. Sensitivity of the automated dipstick reader for detection of glucosuria was 23% for canine samples and 68% for feline samples; specificity was 99% and 98%, respectively. Visual readings were more accurate than automated readings for both canine and feline urine. The UGCR was significantly correlated with absolute urine glucose concentration for both dogs and cats, yet there was incomplete distinction between dipstick categories for glucose concentration and UGCR. Urine dipstick readings for dogs and cats were useful for ruling glucosuria in when the result was positive but not for ruling it out when the result was negative. The evaluated dipsticks were more accurate for detection of glucosuria in cats than in dogs. Visual dipstick readings were more accurate than automated readings. The UGCR did not appear to provide additional useful information.