PURPOSE: This study aimed to determine the best cutoff line for insulin-like growth factor (IGF)-I and insulin-like growth factor binding protein (IGFBP)-3 to discriminate between growth hormone deficiency (GHD) patients and the control group. METHODS: Two hundred thirty subjects with normal controls (129 boys and 101 girls, aged 7-15 years), 14 patients with complete GHD (12 boys and 2 girls), and 17 patients with partial GHD (9 boys and 8 girls) were studied. IGF-I serum concentrations were measured by radioimmunoassay (RI), and IGFBP-3 concentrations were measured by immunoradiometric assay (IRMA). RESULTS: The receiver operating characteristic (ROC) plot analysis showed that the best IGF-I and IGFBP-3 cutoff line was at -1 standard deviation (SD). By comparing IGF-I serum levels of GHD children within 1 SD of normal control, we determined the sensitivity (S) (87.5-100%) and specificity (Sp) (80-84.6%) according to the age group. For IGFBP-3 , we determined the following values: S (58.7-85.7%) and Sp (79.2-85.5%). Eleven of 1 4 patients with complete GHD (78.5%) and 16 of 17 patients with partial GHD (94.1%) had IGF-I concentrations equal to or below -1 SD of the control group mean. Ten of 12 complete GHD children (83.3%) and 13 of 17 partial GHD children (76.5%) had IGFBP-3 concentrations equal or below -1 SD of the control group mean. CONCLUSION: We conclude that the measurement of IGF-I and IGFBP-3 concentrations might provide essential supplementary data in the diagnostic evaluation of patients with GHD. Our results support the need to use cutoff lines based on below -1 SD of the control.