A multicentric retrospective study of serum/plasma urea and creatinine concentrations in dogs using univariate and multivariate decision rules to evaluate diagnostic efficiency

Background: Urea and creatinine are the most frequently used indirect markers in plasma and serum of glomerular filtration rate in dogs. Both have been shown to lack sensitivity but their diagnostic efficiency for the diagnosis of kidney disease has been minimally investigated. Objective: The purpose of this retrospective study was to investigate the influence of possible factors of variation on both analytes and to determine whether specific decision rules should be drawn up for subpopulations of dogs. Methods: The results of urea and creatinine measurements, breed, sex, age, and health status (healthy, renal disease, or nonrenal disease) of 3822 dogs were collected from the archives of 5 veterinary clinics. Data were analyzed with univariate and multivariate decision rules with and without adjustment. Results: There were significant effects and interactions of almost all of the sources of variation. Slight improvements in diagnostic efficiency were obtained by adjusting the decision rules to these sources of variations. Univariate decision rules gave approximately the same diagnostic efficiency for urea and creatinine concentrations, with sensitivity and specificity in the range of 70% and 90%, respectively, using the upper limit of the reference interval as the threshold value. Multivariate decision rules provided only minor improvements in diagnostic efficiency. Conclusion: Simultaneous measurement of both urea and creatinine is of limited diagnostic value over the analysis of a single variable. Creatinine is the preferred analyte as it is affected by fewer extrarenal factors of variation.