Evaluation of endocrine tests. B: screening for hypercortisolism.

Background: While reference values for 24-hour free urinary cortisol excretion and the overnight 1 mg dexa-methasone-suppression test in the healthy population are available, cut-off values in patients clinically suspected of Cushing's syndrome have to be established.

Methods: This was a prospective follow-up study in one academic centre of 144 patients with clinical suspicion of Cushing's syndrome (group A) and 50 patients with adrenal incidentaloma (group B) who were referred for putative hypercortisolism between 1 January 1993 and 1 January 2003. The 24-hour urinary free cortisol and post-dexamethasone plasma cortisol were measured. Accurate diagnosis of (absence of) Cushing's syndrome was confirmed by histopathological data and long-term follow-up. Based on the data obtained in group A, sensitivity, specificity and receiver operating characteristic (ROC) curves were calculated.

Results: Complete follow-up was obtained in 86%, and partial follow-up was obtained in 8% of patients. Median follow-up was 36 (1 to 122) months. In group A, 17 patients were found to have Cushing's syndrome. In this group median 24-hour urinary free cortisol was 77 (<5 to 51458) mmol/24 hours and median post-dexamethasone plasma cortisol was <50 (<50 to 4900) nmol/l. Area under the ROC curve was 0.958 for 24-hour urinary free cortisol and 0.985 for post-dexamethasone plasma cortisol. Optimal cut-off values were 180 nmol/24 hours (sensitivity 94%, specificity 94%) and 95 nmol/l (sensitivity 100%, specificity 94%) respectively.

Conclusion: We established cut-off values for 24-hour free urinary cortisol excretion (180 nmol/24 hours) and for post-dexamethasone plasma cortisol (95 nmol/l) in the evaluation of patients referred for hypercortisolism.