A urine cortisol:creatinine (c:c) ratio, determined from a free-catch morning sample, was evaluated in each of 83 dogs as a screening test for hyper-adrenocorticism. The dogs evaluated were allotted to 3 groups, including 20 healthy dogs, 40 dogs with confirmed hyperadrenocorticism (HAC), and 23 dogs with polyuria and polydipsia not attributable to HAC (polyuria/polydipsia group; PU/PD). Overlap in the urine c:c ratios (mean +/- SEM), comparing results from the healthy dogs (5.7 x 10(-6) +/- 0.9) with those from the HAC dogs (337.7 x 10(-6) +/- 72.0) was not found. However, 11 (64%) of the 18 values from the PU/PD dogs (42.6 x 10(-6) +/- 9.4) were above the lowest ratio in the HAC group and 50% of the HAC group had a urine c:c ratio below the highest value in the PU/PD group. When the mean urine c:c ratio (+/- 2 SD) for the group of healthy dogs was used as a reference range, 100% of the HAC dogs and 18 (77%) of 23 dogs in the PU/PD group had abnormal urine c:c ratios. The sensitivity of the urine c:c ratio to discriminate dogs with HAC was 100%. The specificity of the urine c:c ratio was 22% and its diagnostic accuracy was 76%. On the basis of our findings, a urine c:c ratio within the reference range provides strong evidence to rule out HAC. However, abnormal urine c:c ratios are obtained from dogs with clinical diseases other than HAC. Therefore, measurement of a urine c:c ratio should not be used as the sole screening test to confirm a diagnosis of HAC.