OBJECTIVE: This study investigates whether urinary cortisol (UC) and UC-to-creatinine ratio (UCCR) perform better than basal serum cortisol (BSC) in identifying dogs with hypoadrenocorticism (HA). METHODS: A retrospective, multicenter study with 120 client-owned dogs included: 20 with HA, 42 healthy, and 60 with diseases mimicking HA. The UC and UCCR were determined on urine samples using a chemiluminescent enzyme immunoassay. The diagnostic performance of the UC and UCCR were assessed based on receiver operating characteristics (ROC) curves. RESULTS: A cutoff value of UC <
2 µg/dL revealed 100% sensitivity (95% CI, 83.2 to 100) and 90.0% specificity (95% CI, 79.5 to 96.2) in diagnosing HA. A cutoff value of UCCR <
8.5 X 10-6 revealed 100% sensitivity (95% CI, 83.1 to 100) and 71.7% specificity (95% CI, 58.6 to 82.6) in diagnosing HA. A cutoff value of BSC <
2 µg/dL and <
1 µg/dL revealed 100% sensitivity (95% CI, 83.2 to 100) and 51.7% specificity (95% CI, 38.5 to 64.8) and 100% sensitivity (95% CI, 83.9 to 100) and 90% specificity (95% CI, 79.8 to 95.3) in diagnosing HA, respectively. CONCLUSIONS: BSC <
1 µg/dL showed the same sensitivity but higher specificity than BSC <
2 µg/dL. The UC <
2 µg/dL showed noninferior performance compared with the BSC <
1 µg/dL. CLINICAL RELEVANCE: UC should be considered a promising screening test for canine HA.