Evidence for two subtypes of Cushing's disease based on the analysis of episodic cortisol secretion

To investigate the pathogenetic mechanisms of Cushing's syndrome, we studied variations in plasma cortisol levels (episodic variations, or pulses) over 24 hours in 51 normal subjects, 14 patients with adrenal adenoma, and 46 patients with Cushing's disease. Data were obtained both from our patients and from the literature. As compared with normal subjects, patients with adrenal adenoma had fewer spikes in cortisol levels (defined as an elevation of at least 10 per cent and no less than 1 microgram per deciliter), and the spikes were lower both in absolute terms (4.0 +/- 1.8 vs. 5.1 +/- 2.2 micrograms per deciliter, P less than 0.05) and in terms of the percentage of the preceding trough concentration (23 +/- 7 vs. 123 +/- 74 per cent, P less than 0.001). Patients with Cushing's disease seemed to fall into two groups: those with hypopulsatile and those with hyperpulsatile secretion. The hypopulsatile group had a normal number and absolute height of spikes, but their height relative to the preceding trough concentration was lower than in controls (42 +/- 16 vs. 123 +/- 74 per cent, P less than 0.005). In contrast, the hyperpulsatile group had a similar number of spikes as the hypopulsatile group, but their absolute and relative heights were twice as great (12.7 +/- 2.3 vs. 6.0 +/- 1.6 micrograms per deciliter and 84 +/- 40 vs. 42 +/- 16 per cent, respectively; P less than 0.001 for both). We hypothesize that the Cushing's disease in the second group of patients may have been caused by increased hypothalamic release of, or pituitary responsiveness to, corticotropin-releasing factor, whereas that in the first group may represent pituitary oversecretion of corticotropin that is relatively independent of corticotropin-releasing factor.