Objective: To analyse the diagnostic role of serum IGF-I, IGF-binding protein-3 (IGFBP-3), IGF-I/IGFBP-3 molar ratio and urinary GH (uGH) excretion in adult GH deficiency (GHD). Design: Twenty-seven adults (age range: 18-71 years) with severe GHD, defined by a peak GH response to an insulin tolerance test below 3 μg/l in patients with at least one additional pituitary hypofunction. Reference values were established from a selected age- and body mass index-matched population (154 healthy adults grouped in four age groups). Methods: IGF-I and IGFBP-3 were measured by RIA (Nichols and results expressed as standard deviation (S.D.) scores from our reference population and assay normative data (S.D. score Nichols). uGH was measured by IRMA. Results: Within the control group, IGF-I, IGFBP-3, IGF-I/IGFBP-3 ratio standardisation regarding our control population and IGF-I with respect to the assay normative data resulted in disappearance of age-related differences. However, IGFBP-3 S.D. score Nichols resulted in mean values between +1.4 and +2.5 S.D. score. Greatest diagnostic efficiency was for IGF- I standardised with respect to our controls (97.2%), followed by S.D. score IGFBP-333 (92.9%). S.D. score IGF/IGFBP-3 ratio and uGH showed poor diagnostic efficiency. Any combination of at least two abnormal parameters raised specificity to 100%. IGF-I standardised with respect to assay reference (S.D. score Nichols) showed similar diagnostic value (95.0%) whereas IGFBP-3 showed low sensitivity (33.3%). Within the GHD patients, those with three or more additional deficiencies had lower S.D. score IGF-I than those with only two or one. Conclusion: We underline the importance of an appropriate reference population for correct interpretation of GH secretion markers. Considering our results, specificity obtained with two simultaneous abnormal parameters when referred to an adequate reference population may add valuable information to alternative GH stimulation tests to confirm adult GHD.