Circadian variation in Cushing's disease and pseudo-Cushing states by analysis of F and ACTH pulsatility

Distinguishing pituitary-dependent Cushing's disease from pseudo-Cushing's states can present a diagnostic challenge. Although many studies potentially discriminate between the 2, only the dexamethasone-suppressed corticotropin-releasing hormone (CRF) stimulation test at 15 minutes is 100% sensitive or specific. We measured baseline profiles of F and ACTH in 31 Cushing's disease patients, 11 with pseudo-Cushing's and 17 controls. Venous blood was collected at 30 minute intervals for 24-h. Subjects also had CRF stimulation tests and 2.0 mg/day dexamethasone suppression tests. F and ACTH profiles were analyzed for circadian rhythmicity, variability, and pulsatility. Relative circadian amplitude was decreased in Cushing's disease compared to both pseudo-Cushing's and normal states. Relative pulse amplitude was reduced in Cushing's disease. Because of this dampening of circadian and pulsatile variations, the overall variability of F and ACTH levels around their mean levels as quantified by the intra-series coefficient of variation (CV), was also decreased in Cushing's disease compared to pseudo-Cushing's and normal states. A F 24-h CV<40% was able to distinguish Cushing's disease from pseudo-Cushing's with 100% sensitivity (95% confidence interval (CI), 88-100%) and specificity (CI, 71-100%). An ACTH CV<40% had 97% sensitivity (CI, 83-100%) and 100% specificity (CI, 71-100%). An overnight 8-h F CV <40% also distinguished Cushing's disease from pseudo-Cushing's with 100% sensitivity (CI, 88-100%) and specificity (CI, 71-100%). These data show that a simple index of total temporal variability (the intra-series CV) derived from the analysis of basal F profiles, provides a useful method to distinguish Cushing's disease from pseudo-Cushing's. A F or ACTH CV <40% discriminates Cushing's disease from pseudo-Cushing's and reflects reduced circadian and pulsatile variations.