Basal cortisol test limits for diagnosing Addison's disease in UK dogs

Abstract

INTRODUCTION: Canine hypoadrenocorticism (HA) is characterized by glucocorticoid (and often mineralocorticoid) deficiency and typically requires ACTH stimulation testing for diagnosis. A basal (pre-ACTH) cortisol >55 nmol/L is widely used to exclude HA, but prior studies were limited to referral populations. Its utility in first-opinion practice remains unclear. OBJECTIVES: To evaluate the diagnostic performance of basal cortisol for identifying HA, and its added value alongside post-ACTH cortisol, using UK laboratory data predominantly from first-opinion practices and the remainder from referral centers. MATERIALS AND METHODS: Retrospective analysis of 1,017 ACTH stimulation tests (January 2019-April 2023) from a UK veterinary diagnostic laboratory. After excluding cases tested for hypercortisolism, 878 cases remained: 170 from a referral center (RC) with full clinical history and 708 predominantly from first-opinion (FO) practices. Serum cortisol was measured by radioimmunoassay. HA was defined as post-ACTH cortisol ≤55 nmol/L. Diagnostic performance metrics were calculated for basal cortisol cut-offs of ≤55 and ≤22 nmol/L. RESULTS: HA prevalence was 8.4% in RC, 4.4% in FO, and 5.1% in combined groups (RC + FC). RC group basal cortisol ≤55 nmol/L showed a high sensitivity (93%) and NPV (99%), but low specificity (77%), and PPV (25%). Reducing the cut-off to ≤22 nmol/L improved specificity (92%) and PPV (50%), maintaining sensitivity (93%) and NPV (99%). However, when assessing the FO group only or when combined with RC (RC + FO), the ≤22 nmol/L cut-off sensitivity was slightly reduced (90 and 91%, respectively). Receiver operating characteristic analysis yielded areas under the curve of 0.93-0.97 for basal cortisol and 1.0 for post-ACTH cortisol across all datasets. Pre-ACTH cortisol provided no additional diagnostic value when post-ACTH cortisol was known. CONCLUSION: Basal cortisol >55 nmol/L is a reliable rule-out test for canine HA in both referral and first-opinion settings, at prevalences lower than seen in referral practices. However, when an ACTH stimulation test is conducted, post-ACTH cortisol alone provides perfect diagnostic accuracy, rendering pre-ACTH cortisol redundant. These findings support omitting pre-ACTH cortisol in routine ACTH stimulation testing to streamline diagnostics and reduce costs without compromising diagnostic performance.