Objective: The diagnosis of Cushing's syndrome remains a major challenge in clinical endocrinology. Various screening tests are commonly used to support a biochemical diagnosis in the context of clinical suspicion. The aim of this study was to compare the sensitivity in the diagnosis of Cushing's syndrome of a single in-patient sleeping midnight cortisol to a standard 48-hour in-patient low-dose dexamethasone suppression test (LDDST) during the same admission.
Design: A retrospective analysis was performed on 150 patients investigated in our department between the years 1970 and 1994 with a confirmed diagnosis of Cushing's syndrome.
Patients: One hundred and fifty patients with a diagnosis of Cushing's syndrome were analysed: 110 with Cushing's disease; 12 with tumours with ectopic ACTH secretion; 8 with ACTH dependent Cushing's syndrome of so far undetermined origin; 17 with cortisol secreting adrenal tumours; 3 with adrenocortical nodular hyperplasia. Twenty normal volunteers and nine patients with non-endocrine conditions were also investigated as controls.
Measurements: Plasma cortisol was measured by radioimmunoassay (RIA) in the 122 patients presenting after 1980, and by fluorimetry prior to this date.
Results: In all the control subjects the sleeping midnight cortisol was < 50 nmol/l, below the lowest standard of the routine in-house RIA. In every patient with Cushing's syndrome the sleeping midnight cortisol was detectable with a value greater than 50 nmol/l, with a range of 70-2000 nmol/l. In contrast, in three cases, all of whom had proven Cushing's disease on histology, there was uncharacteristic complete suppression of plasma cortisol to < 50 nmol/l following the LDDST.
Conclusion: In this series of 150 cases, a single in-patient sleeping midnight cortisol above 50 nmol/l had a 100% sensitivity for the diagnosis of Cushing's syndrome, clearly different from normal subjects. In contrast, the low-dose dexamethasone suppression test had a sensitivity of 98% even when the drug was administered as an in-patient. We recommend that a low-dose dexamethasone suppression test should not be used alone for confirmation of Cushing's syndrome since it may miss 2% of cases.