GH deficiency (GHD) in adults has to be shown by a single provocative test, provided that it is validated. Insulin tolerance test (ITT) has been indicated as the test of choice; now also glucagon test is validated and represents an alternative. The GHRH plus arginine (ARG) test and testing with GHRH plus a GH secretagogue are equally reliable diagnostic tools, and are now considered as 'golden' standards as ITT. Childhood-onset (CO) GHD needs retesting in late adolescence or young adulthood; this is a major clinical challenge and raises questions about the most appropriate method and cut-off value. Appropriate re-evaluation of GH status is represented by simple measurement of IGF1 concentration off rhGH treatment. Clearly, low IGF1 levels are evidence of persistent severe GHD in subjects with genetic GHD or panhypopituitarism. However, normal IGF1 levels never rule out severe GHD and CO-GHD with normal IGF1 levels must undergo a provocative test. The appropriate GH cut-off limit is specific for each provocative test. As shown by the ROC curve analysis, in late adolescents and young adults, the lowest normal GH peak response to ITT is 6.1 microg/l while that to GHRH+ARG test is 19.0 microg/l. These cut-off limits, however, are just indicative as being variable as a function of the assay used. No other test is validated for retesting. As GHRH+ARG test mostly explores the GH-releasable pool, normal GH response would be verified by a second ITT in order to rule out subtle hypothalamic defect.