Urinary cortisol test has limited accuracy for diagnosing dog

Abstract

OBJECTIVE: To establish the de novo reference interval (RI) for urinary cortisol-to-creatinine ratio (UCCR) in healthy dogs (HDs) using the currently available chemiluminescent enzyme immunoassay antibody (Veterinary Cortisol; IMMULITE 2000 XPi; Siemens Healthineers) and to evaluate UCCR diagnostic performance in dogs with hypercortisolism (HC). METHODS: This was a retrospective, single-center, observational study. Stored urine samples from dogs with HC or diseases mimicking HC (DMHC) at the time of diagnosis were selected from July 2019 through November 2022. Healthy dogs were prospectively included. The diagnostic performance of the UCCR was assessed based on receiver operating characteristic curves. RESULTS: One hundred seventy-two dogs were included: 97 with HC, 35 with DMHC, and 40 HDs. The de novo RI for UCCR in HDs was between 3 X 10-6 (90% CI, 2.3 X 10-6 to 3.8 X 10-6) and 26 X 10-6 (90% CI, 29.7 X 10-6 to 35.0 X 10-6). The median UCCR was significantly higher in dogs with HC (70.9 X 10-6; 6.8 X 10-6 to 882.2 X 10-6) as compared to dogs with DMHC (15 X 10-6; 2.63 X 10-6 to 137.8 X 10-6) and HDs (9.1 X 10-6; 3.9 X 10-6 to 36.3 X 10-6). The area under the receiver operating characteristic curve for UCCR to differentiate HC dogs from dogs with DMHC was 0.85 (95% CI, 0.78 to 0.92). Using the upper limit of the de novo RI as the cutoff value (UCCR > 26 X 10-6), the sensitivity and the specificity for the UCCR in diagnosing HC were 80.4% (95% CI, 71.1% to 87.8%) and 71.4% (95% CI, 53.7% to 85.4%), respectively. CONCLUSIONS: Using the upper limit of the de novo RI, UCCR showed modest performances not only due to low specificity but also due to a sensitivity of only 80.4%. CLINICAL RELEVANCE: UCCR should not be used alone to rule out HC in dogs, and, when the clinical suspicion for HC is present, other endocrine tests should be pursued.