The reliability and reproducibility of provocative stimuli of growth hormone (GH) secretion in the diagnosis of GH deficiency are still controversial both in childhood and in adulthood. The combined administration of GH-releasing hormone (GHRH) and arginine (ARG), which likely acts via inhibition of hypothalamic somatostatin release, is one of the most potent stimuli known so far and has been proposed recently as the best test to explore the maximal somatotrope capacity of somatotrope cells. However, it is well known that, usually, provocative stimuli of GH secretion suffer from poor reproducibility and that of the GHRH + ARG test has still to be verified. We aimed to verify the between- and within-subject variability of the GH response to the GHRH + ARG test in normal subjects during their lifespan as well as in hypopituitaric patients with GH deficiency (GHD). In 10 normal children (C: six male and four female, age 12.3 +/- 0.9 years, body mass index (BMI) = 16.6 +/- 0.7 kg/m2, pubertal stages I-III), 18 normal young adults (Y: ten male and eight female, age 31.1 +/- 1.3 years, BMI = 21.4 +/- 0.4 kg/m2), 12 normal elderly subjects (E: two male and ten female, age 74.4 +/- 1.8 years, BMI= 22.6 +/- 0.6 kg/m2) and 15 panhypopituitaric GH-deficient patients (GHD: nine male and six female, age 40.9 +/- 4.1 years, BMI= 22.7 +/- 1.0 kg/m2), we studied the inter- and intra-individual variability of the GH response to GHRH (1 microg/kg i.v.) + ARG (0.5 g/kg i.v.) in two different sessions at least 3 days apart. The GH responses to GHRH + ARG in C (1st vs 2nd session: 61.6 +/- 8.1 vs 66.5 +/- 9.4 microg/l), Y (70.4 +/- 10.1 vs 76.2 10.7 microg/l) and E (57.9 14.8 vs 52.1 +/- 8.0 microg/l) were similar and reproducible in all groups. The somatotrope responsiveness to GHRH + ARG also showed a limited within-subject variability (r = 0.71, 0.90 and 0.89 and p < 0.02, 0.0005 and 0.0005 for C, Y and E, respectively). Similarly in GHD, the GH response to the GHRH + ARG test showed a good inter- (1st vs 2nd session: 2.3 +/- 0.5 vs 2.2 +/- 0.6 microg/l) and intra-individual reproducibility (r = 0.70, p < 0.005). The GHRH + ARG-induced GH responses in GHD were markedly lower (p < 0.0005) than those in age-matched controls and no overlap was found between GH peak responses in GHD and normal subjects. In normal subjects, the GH response to GHRH + ARG is very marked, independent of age and shows limited inter- and intra-individual variability. The GH response to the GHRH + ARG test is strikingly reduced in panhypopituitaric patients with GHD, in whom the low somatotrope responsiveness is reproducible. Thus, these findings strengthen the hypothesis that GHRH + ARG should be considered the most reliable test to evaluate the maximal secretory capacity of somatotrope cells and to distinguish normal subjects from GHD patients in adulthood.