The metyrapone and dexamethasone suppression tests for the differential diagnosis of the adrenocorticotropin-dependent Cushing syndrome: a comparison.

OBJECTIVE: To develop criteria for interpreting results of the metyrapone test for the differential diagnosis of the adrenocorticotropin (ACTH)-dependent Cushing syndrome and to compare its diagnostic accuracy with that of the high-dose dexamethasone suppression test.

DESIGN: Retrospective cohort study.

SETTING: Inpatient research ward.

PATIENTS: 186 patients believed to have the ACTH-dependent Cushing syndrome who had the metyrapone test, the dexamethasone test, and a surgical procedure to remove the source of excessive ACTH.

MEASUREMENTS: The sensitivity, specificity, and diagnostic accuracy were determined for the metyrapone test using urine excretion of hydroxysteroid and plasma levels of 11-deoxycortisol. For the dexamethasone suppression test, urine excretions of 17-hydroxysteroid (17-OHS) and free cortisol were used.

MAIN RESULTS: 156 patients had pituitary disease, 15 had ectopic ACTH secretion, and 15 had no diagnosis after pituitary surgery. Of those 15 patients, 14 were ultimately classified as having pituitary disease on the basis of follow-up, and 1 was found to have ACTH-independent Cushing syndrome. After administration of metyrapone, stimulation of 17-OHS excretion greater than 70% or of a plasma 11-deoxycortisol level greater than 400-fold did not result in the misclassification of any of the patients with surgically confirmed cases of ectopic ACTH secretion. When these criteria were combined, the percentage of correct predictions (122 of 186 [71%; 95% CI, 62% to 79%]) was higher than that for either steroid alone (116 of 186 [62%; CI, 52% to 71%]) for excretion of 17-OHS and that for plasma 11-deoxycortisol (82 of 186 [44%; CI, 34% to 54%]). When the criteria for both the metyrapone test and the dexamethasone test were combined, the percentage of correct predictions (153 of 186 [82%; CI, 75% to 87%]) was higher than that obtained when the criteria for either test alone were used (P = 0.001). Similar results were found when the 15 patients with indeterminate surgery were assigned to the appropriate group on the basis of follow-up. When the criteria for both the metyrapone and dexamethasone tests were combined to identify patients with the pituitary Cushing syndrome, the sensitivity and diagnostic accuracy improved to 88% and 89%, respectively.

CONCLUSIONS: The metyrapone test, which can be done in 48 hours, has a sensitivity and specificity for the diagnosis of the Cushing syndrome identical to that of the standard 6-day high-dose dexamethasone suppression test. Combining both tests results in greater accuracy than that obtained with either test alone.