Divergence between growth hormone responses to insulin-induced hypoglycaemia and growth hormone-releasing hormone in patients with non-functioning pituitary macroadenomas and hyperprolactinaemia

OBJECTIVE The GH responses to the insulin tolerance test (ITT) and growth hormone-releasing hormone (GHRH) may yield different results in patients with pituitary lesions. The GH responses to these stimuli were compared in patients with untreated non-functioning pituitary macroadenomas, who represent an important cause of GH deficiency.DESIGN Analysis of peak on to ITT and to 100 mu g GHRH in relation to an elevated PRL level (> 200 mIU/l for males and > 600 mIU/l for females) as an indication of hypothalamic-pituitary dysregulation, as well as in relation to other anterior pituitary hormone deficiencies. A peak GH <5 mu g/l in either test indicated GH deficiency.PATIENTS Twenty females and 14 males (median age 52 (23-77) years) evaluated preoperatively in a university hospital setting.RESULTS In the whole group the median peak GH to GHRH (3.6 (0.9-26.3) mu g/l) was higher than to ITT (1.6 (0.2-7.8) mu g/l, P <0.001). This difference was seen only in 19 patients with concomitant hyperprolactinaemia (P <0.001). When hyperprolactinaemia was present, an insufficient GH peak was demonstrated by ITT in 16 cases and by GHRH stimulation in 7 cases (P <0.01). The frequency of an insufficient GH peak by ITT (13 cases) and by GHRH (14 cases) was similar in the normoprolactinaemic patients. In addition, 9 of 10 patients with an impaired response to ITT and a normal response to GHRH were hyperprolactinaemic compared to 7 of 19 patients with GH deficiency as assessed by both stimuli (P <0.02). Peak GH to ITT was lower in 24 patients with, compared to 10 patients without, other hormonal deficiencies (1.4 (0.2-5.6) vs 3.0 (1.0-7.8) mu g/l, P <0.02), but was not related to elevated PRL. In contrast, GHRH-stimulated GH was higher in hyperprolactinaemic than in normoprolactinaemic patients (5.9 (1.6-26.3) vs 2.9 (0.9-5.4) mu g/l, P <0.001) and was not related to the presence of other pituitary hormone deficiencies. Analysis of covariance confirmed that peak GH to ITT was negatively associated with the presence of other pituitary hormone deficiencies (P <0.01), whereas peak GH to GHRH was positively related to an elevated PRL level (P <0.02). nasal GH was positively correlated with PRL (R(s) = 0.36, P <0.05).CONCLUSIONS This study demonstrates that ITT and GHRH tests cannot be used interchangeably in diagnosing GH deficiency in patients with non-functioning pituitary macroadenoma and hyperprolactinaemia. If the ITT is considered to be the reference test, GH deficiency as assessed by GHRH can be missed in patients with hyperprolactinaemia. This disparity is probably due to a different mechanism of action of these stimuli. Hyperprolactinaemia may be associated with a diminished somatostatin tone, leading to a higher basal and GHRH-stimulated GH, without having an effect on peak GH to ITT.