A single midnight serum cortisol measurement distinguishes Cushing's syndrome from pseudo-Cushing states

Cushing's syndrome (CS) may be difficult to distinguish from pseudo-Cushing states (PCS) based on physical findings or urinary glucocorticoid excretion. As the lack of diurnal variation in serum cortisol is characteristic of CS, we studied whether diurnal cortisol determinations could discriminate CS from PCS. Two hundred and sixty-three patients were evaluated: 240 had CS, and 23 had PCS. Urine was collected for 24 h for measurement of cortisol and 17-hydroxycorticosteroids (17OHCS). Blood was drawn at 2300, 2330, 0000, 0030, and 0100 h and at 0600, 0630, 0700, 0730, and 0800 h the next morning for serum cortisol determination. The main outcome measure was the sensitivity of these parameters for the diagnosis of CS at 100% specificity. A midnight cortisol value greater than 7.5 microg/dL correctly identified 225 of 234 patients with CS and all PCS patients. This sensitivity (96%) was superior to that obtained for any other measure, including urinary cortisol (45%), 17OHCS (22%), any other individual cortisol time point (10-92%), the morning (23%) or the evening (93%) cortisol mean, and the ratio (11%) of morning to evening values. We conclude that at 100% specificity, a single serum cortisol value above 7.5 microg/dL at midnight discriminates CS from PCS with higher sensitivity than 24-h urinary cortisol or 17OHCS, or other individual or combined measures of serum cortisol.