Background: Re-testing for GH secretion is needed to confirm the diagnosis of GH deficiency (GHD) after adult height achievement in childhood-onset GHD (COGHD). Aim: To define the cut-off of GH peak after retesting with GH-releasing hormone plus arginine (GHRHarg) in the diagnosis of permanent GHD in COGHD of different etiology. Patients and methods: Eighty-eight COGHD (median age 17.2 y), 29 idiopathic GHD (IGHD), 44 cancer survivors (TGHD) and 15 congenital GHD (CGHD) were enrolled in the study; 54 had isolated GHD (iGHD) and 34 had multiple pituitary hormone deficiencies (MPHD). All were tested with insulin tolerance test (ITT) and GHRHarg. IGHD with a GH response to ITT ≥6μg/L were considered true negatives and served as the control group, and patients with a GH response <6μg/L as true positives. Baseline IGF-I was also measured. The diagnostic accuracy of GHRHarg testing and of IGF-I SDS in patients with GHD of different etiologies was evaluated by ROC analysis. Results: Forty-six subjects with a GH peak to ITT ≥6μg/L and 42 with GH peak <6 μg/L showed a GH peak after GHRHarg between 8.8-124μg/L and 0.3-26.3μg/L, respectively; 29 IGHD were true negatives, 42 were true positives and 17 with a high likelihood GHD showed a GH peak to ITT ≥6μg/L. ROC analysis based on the etiology indicated the best diagnostic accuracy for peak GH cutoffs after GHRHarg of 25.3 μg/L in CGHD, 15.7 in TGHD, and 13.8 in MPHD, and for IGF-1 SDS at -2.1 in CGHD, -1.5 in TGHD, and -1.9 in MPHD. Conclusions: Our findings indicate that the best cut-off for GH peak after retesting with GHRHarg changes according to the etiology of GHD during the transition age. Based on these results the diagnostic accuracy of GHRHarg remains questionable.