Newer Generation Immuno-Assays in Screening for Adrenal Insufficiency: comparison of baseline morning cortisol levels with cosyntropin stimulation testing

Abstract Background/Objective Serum cortisol immunoassays with improved specificity for cortisol have prompted revisiting the definition of a normal response at 60-minute cosyntropin stimulation testing (CST-60) to ≥14 μg/dL. European and American Endocrine Societies define values of AM-cortisol >10 μg/dL as adequate to rule out adrenal insufficiency (AI) in patients with long-term use of corticosteroids. We sought to compare AM-cortisol levels with CST-60 in patients suspected of AI from various causes. Methods A retrospective review was performed from the electronic health record of a large midwestern hospital in patients who were evaluated for AI. Results There were 95 patients (51women, age 58.8±17.8 SD years) with 43 tests performed outpatient, 48 inpatient, and 4 ICU settings. The pre-test diagnoses were adrenal abnormalities (n=10), hyponatremia (n=11), hypoglycemia (n=9), fatigue (n=19), hypotension (n=35), pituitary abnormality (n=7) and shock (n=4). Binary logistic regression analyses (BLRA) and receiver operating curves (ROC) compared AM-cortisol as predictors of a normal CST-60 ≥ 14.6 μg/dL and to the previous criteria of a CST-60 ≥18.0 μg/dL. Receiver operating curves showed that an AM-cortisol of 11.1 μg/dL had a specificity of 93% for a CST-60 >14.6 g/dL. Using historical levels of CST-60 ≥ 18.0 μg/dL there would be 17% false positives requiring unnecessary dynamic testing. Conclusions In this community setting with low risk of AI (16%), a baseline AM-cortisol of 11.1 μg/dL can reliably preclude the need for a CST-60. Clinicians should be aware of the new cutoffs for the assays available to them when evaluating patients for AI.