Plasma cortisol cut-off concentrations using the dexamethasone suppression test in patients with adrenal incidentalomas.
Jørgensen Nanna Thurmann NT, Borresen Stina Willemoes SW, Hilsted Linda L, Christoffersen Christina C et al.
Assessment of cortisol secretion in patients with adrenal incidentalomas remains challenging. We evaluated the guideline-recommended cortisol cut-off value for cortisol suppression (50 nmol/L) following a 1 mg overnight dexamethasone suppression test (DST) against a control-derived threshold and compared patient classification using the two thresholds. Secondary objectives included assessing associations of pre-DST ACTH, post-DST dexamethasone, and comorbidities with post-DST cortisol. In this cross-sectional study (NCT07350031), pre- and post-DST plasma samples from patients with adrenal incidentalomas (N = 108) and matched controls (N = 101) were prospectively collected, and cortisol was analysed using immunoassay (Elecsys®Cort II) and LC-MS/MS. Post-DST dexamethasone ≥3 nmol/L defined appropriate dexamethasone exposure. The 97.5th percentile of LC-MS/MS-measured post-DST cortisol in controls defined the control-derived threshold. Associations between pre-DST ACTH, post-DST dexamethasone, and post-DST cortisol were assessed. The performance of pre-DST ACTH for identifying post-DST cortisol >50 nmol/L was evaluated using receiver operating characteristic analysis. Logistic regression examined associations between post-DST cortisol and diabetes, dyslipidaemia, hypertension, and osteoporosis. The 97.5th percentile of post-DST cortisol in controls was 66 nmol/L. The 50 nmol/L threshold corresponded to the 90th percentile of controls and classified 44% of patients as having hypercortisolism, compared with 26% using the control-derived threshold (66 nmol/L). Pre-DST ACTH showed poor discrimination for post-DST cortisol >50 nmol/L (AUC 0.69). Post-DST cortisol was not associated with dexamethasone concentrations or comorbidities. The guideline-recommended post-DST cortisol threshold of 50 nmol/L may be too low, suggesting that higher thresholds warrant further evaluation.