The acute administration of glucocorticoids is a new stimulus of growth hormone (GH) secretion in man. In order to ascertain its point of action, and also the suitability of this new test as a diagnostic tool in GH pathological states, 33 subjects were studied. Eight of them were normal controls, and 25 were patients with tumors affecting the hypothalamopituitary area. A glucocorticoid stimulus, dexamethasone 4 mg i.v. was administered at 0 min and GH levels (means ± SEM, µg/l) were measured during the following 5 h. In addition, GH-releasing hormone (GHRH) and clonidine were employed as either pituitary or hypothalamic GH stimuli. Dexamethasone administration to normal subjects did not alter GH levels in the first 2 h of the test. Afterwards, a GH peak was observed around the third hour, GH levels returning to basal ones thereafter. The dexamethasone-induced GH peak (6.7 ± 1.5) and area under the curve (526 ± 137) were lower than after GHRH (14.0 ± 4.5 and 1,070 ± 369, respectively). In the 14 acromegalic patients studied, the GHRH-induced GH net increase was similar to that observed in controls, while the placebo did not alter GH basal levels. An absence of hypothalamic control was evident because clonidine did not stimulate GH release. On the other hand, and contrary to normal subjects, dexamethasone strongly inhibited GH secretion, the values being significantly lower when calculated either as mean GH peak, or maximum GH increment (Δ). The Δ GH was -2.5 ± 3.1 after placebo, +3.7 ± 4.5 after clonidine, +17.0 ± 3.3 after GHRH and -13.4 ± 4.5 following dexamethasone administration. In 4 patients harboring pituitary prolactinomas, dexamethasone stimulated GH secretion in 1 subject with a microadenoma but was devoid of action in the 3 marcoprolactinoma patients, which had suprasellar extension and absent clonidine-induced GH secretion. Similarly, in 4 out of 5 patients with nonsecreting pituitary adenomas that had a functional pituitary stalk section (suprasellar extension, mild prolactin levels and no GH response to hypothalamic stimulus), dexamethasone was ineffective, while it stimulated GH secretion in the other patient of this group with a normal hypothalamopituitary connection. In addition, in 2 other patients with hypothalamic germinoma and well-preserved GH response to GHRH, dexamethasone was devoid of action. These results indicate that GH stimulation by glucocorticoid is only observed in normal subjects and requires a normal hypothalamopituitary connection. On the contrary, in acromegaly, glucocorticoids induced, instead, an inhibition of GH secretion. In conclusion, glucocorticoid stimulatory action over GH secretion seems to be exerted at the hypothalamic level. On the contrary, in acromegalic patients a ‘paradoxical’ inhibition of GH secretion is observed after glucocorticoid injection. Acute administration of glucocorticoids may be a suitable test in the diagnosis and follow-up of GH pathological states.