A Predictive Risk Score to Diagnose Adrenal Insufficiency in Outpatients: A 7 Year Retrospective Cohort Study

Background: The diagnosis of adrenal insufficiency (AI) requires dynamic tests which may not be available in some institutions. This study aimed to develop a predictive risk score to help diagnose AI in outpatients with indeterminate serum cortisol levels. Methods: Five hundred and seven patients with intermediate serum cortisol levels (3–17.9 µg/dL) who had undergone ACTH (adrenocorticotropin) stimulation tests were included in the study. A predictive risk score was created using significant predictive factors identified by multivariable analysis using Poisson regression clustered by ACTH dose. Results: The seven predictive factors used in the development of a predictive model with their assigned scores are as follows: chronic kidney disease (9.0), Cushingoid appearance in exogenous steroid use (12.0), nausea and/or vomiting (6.0), fatigue (2.0), basal cortisol <9 µg/dL (12.5), cholesterol <150 mg/dL (2.5) and sodium <135 mEq/L (1.0). Predictive risk scores range from 0–50.0. A high risk level (scores of 19.5–50.0) indicates a higher possibility of having AI (positive likelihood ratio (LR+) = 11.75), while a low risk level (scores of <19.0) indicates a lower chance of having AI (LR+ = 0.09). The predictive performance of the scoring system was 0.82 based on the area under the curve. Conclusions: This predictive risk score can help to determine the probability of AI and can be used as a guide to determine which patients need treatment for AI and which require dynamic tests to confirm AI.