Diagnostic Value of Key Clinical Characteristics and Baseline Cortisol in Assessing Adrenal Function in Patients Receiving Glucocorticoid Therapy

Background: Prolonged glucocorticoid therapy may result in multi-organ complications and adrenal insufficiency (AI). To mitigate these risks, clinicians typically taper and discontinue treatment once the underlying disease is controlled. Prior to withdrawal, adrenal function is assessed through dynamic testing. However, dynamic testing often requires multiple blood samples for cortisol measurement, continuous patient monitoring for several hours, and close supervision by medical staff throughout the procedure to assess potential adverse effects. In addition, dynamic tests are typically costly and not all healthcare facilities are equipped to perform such procedures. Objective: Evaluate the diagnostic performance of baseline cortisol and clinical parameters in the assessment of adrenal insufficiency. Methodology: Cross-sectional study of 96 patients on prolonged glucocorticoid treatment at Nguyen Tri Phuong hospital. Insulin tolerance test (IST) was the gold standard for diagnosis of adrenal insufficiency. Results: <0.001). Baseline cortisol had good diagnostic performance, with areas under the receiver operating characteristic curve (AUROC) of 0.83 [95% confidence interval (CI): 0.78-0.89]. If the optimal cutoff value for morning serum cortisol is set at 9.0 µg/dL, the test demonstrates a sensitivity of 69% and a specificity of 67%. We observed that the morning serum cortisol cutoff of 3 µg/dL had a positive predictive value (PPV) of 100% and the morning serum cortisol cutoff of 15 µg/dL had a negative predictive value (NPV) of 100%. Among the predictive models for AI, the model incorporating baseline characteristics such as Cushingoid appearance, duration of glucocorticoid therapy, history of hypertension, and baseline cortisol demonstrated the highest predictive accuracy, with an AUC of 0.87 (95% CI: 0.80-0.94). Conclusion: The assessment of baseline cortisol in conjunction with baseline characteristics offers a reliable method for predicting the risk of adrenal insufficiency, thereby minimizing the necessity for expensive and intricate dynamic testing. In circumstances where dynamic testing is not feasible due to limitations in resources, cost, or expertise, we advocate for the implementation of a model that incorporates baseline characteristics and cortisol levels. This approach provides a practical, cost-effective solution for evaluating adrenal function, requiring only a single blood sample and ensuring broader accessibility.