Cut-off values of urinary free cortisol and cortisone for diagnosing neoplastic hypercortisolism

Quantification of urinary free cortisol excretion is a recommended initial screening test for diagnosing hypercortisolism. This study aims to determine optimal cut-off values for urinary free cortisol (Cortisol_urine) and urinary free cortisone (Cortisone_urine) for diagnosing neoplastic hypercortisolism and to determine whether either one has superior diagnostic accuracy.

We performed a retrospective study using clinical and laboratory data of subjects screened for neoplastic hypercortisolism at Amsterdam UMC between December 2015 and February 2022. Clinical follow-up data was used as the gold standard to define subjects with and without neoplastic hypercortisolism. Subjects collected urine for 24 h on two consecutive days. Cortisol_urine and Cortisone_urine concentrations were measured using LC-MS/MS. The average concentration of both urine collections was used for analysis. Diagnostic accuracy was evaluated by ROC-curve analysis.

Cortisol_urine was available from 27 subjects with, and 417 subjects without neoplastic hypercortisolism. In a subgroup of 17 subjects with, and 292 subjects without neoplastic hypercortisolism, Cortisone_urine was also available. The optimal cut-off value for Cortisol_urine was 120 nmol/24 h (sensitivity 89%, specificity 86%; ROC AUC: 0.937) and 300 nmol/24 h for Cortisone_urine (sensitivity 94%, specificity 88%; ROC AUC: 0.941). The optimal cut-off value for the sum Cortisol_urine + Cortisone_urine was 460 nmol/24 h (sensitivity 94%, specificity 93%; ROC AUC: 0.947).

We established cut-off values for Cortisol_urine and Cortisone_urine for the diagnosis of neoplastic hypercortisolism. Cortisol_urine, Cortisone_urine and their sum had comparable diagnostic accuracies. The proposed cut-off values can be used by other laboratories using LC-MS/MS methods following analytical method comparison and validation of the proposed cut-off values in a different cohort.