Urinary-free cortisol-based thresholds for differentiating ACTH-dependent Cushing: a Spanish validation study

CONTEXT: Differentiating ectopic ACTH secretion (EAS) from Cushing's disease (CD) remains one of the most challenging steps in the diagnostic workup of ACTH-dependent Cushing's syndrome (CS). Urinary free cortisol (UFC) expressed as times above the upper limit of normal (ULN) has been proposed as a simple, noninvasive discriminator, but external validation in independent populations is lacking. OBJECTIVE: To validate the diagnostic performance of UFC×ULN for distinguishing EAS from CD and to explore complementary biochemical markers, including late-night salivary cortisol (LNSC×ULN) and hypokalemia. DESIGN, SETTING, AND PARTICIPANTS: Multicenter retrospective study from the Spanish Cushing Registry including 269 patients with ACTH-dependent Cushing's syndrome (208 CD, 61 EAS) diagnosed and managed in tertiary referral centers. MAIN OUTCOME MEASURES: Diagnostic accuracy of UFC×ULN and LNSC×ULN for discriminating EAS from CD, expressed as area under the ROC curve (AUC), sensitivity, specificity, and predictive value. RESULTS: EAS patients were older (median 59.0 vs 44.9 years; P<0.001) and showed higher UFC×ULN (16.6 vs 3.6; P<0.001) and LNSC×ULN (9.3 vs 1.5; P<0.001). UFC×ULN and LNSC×ULN achieved excellent discriminative performance (AUC 0.90 and 0.92). No EAS occurred with UFC×ULN < 3×ULN, while 40.5% of patients with UFC ≥ 10×ULN had EAS. The combination of severe hypercortisolism (UFC ≥ 10×ULN and LNSC ≥ 9×ULN) plus hypokalemia identified 75% of EAS with 98% specificity. CONCLUSIONS: UFC×ULN thresholds reliably stratify the probability of EAS versus CD. Severe hypercortisolism and hypokalemia strongly predict EAS, supporting a pragmatic diagnostic approach that prioritizes whole-body imaging in high-risk patients and pituitary-centered evaluation in mild cases.