SAT-431 Newer Generation Immuno-Assays in Screening for Adrenal Insufficiency: Comparison of Baseline Morning Cortisol Levels with Cosyntropin Stimulation Testing

Abstract Disclosure: E. Shrestha: None. S.G. Albert: None. A. Ghimire: None. I. Iqbal: None. R.O. Akande: None. S. Portillo Canales: None. Advancements in immunoassays have improved the specificity of serum cortisol measurements compared to older assays. These newer assays allowed revisions of cutoff values for cortisol at 60 minutes for a normal cosyntropin stimulation test (CST), and thus rule out adrenal insufficiency (AI), from cortisol values of 18ug/dL to 14 ug/dL. Similarly, The Endocrine Society 2024 has used these newer assays to define values of AM cortisol which would predict a normal CST in individuals who were on long term corticosteroid therapy and therefore would eliminate the need for CST testing. However, AM cortisol cutoffs to predict a normal CST due to other presumed causes for AI needs reassessments. We aimed to compare morning cortisol levels with CST outcomes, using the newer generation assays cut off values to better predict normal CST responses in patients suspected of AI from various causes. We surveyed 100 patients who had CST between dates from 1/1/2024 - 6/30/2024 from the electronic health record of a midwestern community hospital system. Of the 95 patients, 51 were women, the mean age was 58.8 (SD ±17.8 years), 43 tests were performed as outpatient, 48 tests as inpatient and 4 in ICU settings. The pretest diagnoses were presumed adrenal abnormalities (n=10), hyponatremia (n=10), hypoglycemia (n=9), fatigue (n=19), hypotension (n=35), pituitary abnormality (n=7) and shock (n=4). Receiver operating curves (ROC) for an AM cortisol as a predictor of having a normal response at 60 minutes from a CST using the newer assay definition of a CST>14ug/dL was compared to previous criteria of a CST>18ug/dL. The newer CST assay criteria (>14 ug/dL) had a higher area under the curve (AUC 0.893) when compared to the previous definition of >18ug/dL (AUC = 0.777), p < 0.001. In these analyses, an AM cortisol level of 11.3 µg/dL with the newer assays was shown to preclude the need for CST testing, while with the older cut off criteria of >18ug/dL, an AM cortisol >15 µg/dL would have been required. A logistic regression analysis of the AM cortisol to predict a normal response of CST>14ug/dL showed that a baseline cortisol of 8.8ug/dL had a 95% positive predictive value (PPV). However, the PPV of having AI was less well defined because of the low prevalence of AI in this population. In conclusion, in this community hospital setting with low risk of AI (15%), a baseline AM cortisol between 8.8 µg/dL and 11.1 µg/dL can reliably preclude the need for further cosyntropin stimulation testing to predict a normal CST cortisol >14 µg/dL. The baseline cortisol was less able to predict AI. Further studies in higher risk patients need to be performed to confirm these observations. Presentation: Saturday, July 12, 2025