The appropriate management of GH-deficient patients during transition from childhood to adulthood has not been reported in controlled trials, even though there is evidence to suggest that this phase is associated with specific problems in relation to GH sensitivity. An issue of particular interest is the impact of GH substitution on insulin sensitivity, which normally declines during puberty. We, therefore, evaluated insulin sensitivity (euglycemic glucose clamp) and substrate metabolism in 18 GH-deficient patients (6 females and 12 males; age, 20 +/- 1 yr; body mass index, 25 +/- 1 kg/m2) in a placebo-controlled, parallel study. Measurements were made at baseline, where all patients were on their regular GH replacement, after 12 months of either continued GH (0.018 +/- 0.001 mg/kg day) or placebo, and finally after 12 months of open phase GH therapy (0.016 mg/kg x day). Before study entry GH deficiency was reconfirmed by a stimulation test. During the double-blind phase, insulin sensitivity and fat mass tended to increase in the placebo group [deltaM-value (mg/kg x min), -0.7 +/- 1.1 (GH) vs. 1.3 +/- 0.8 (placebo), P = 0.18; deltaTBF (kg), 0.9 +/- 1.2 (GH) vs. 4.4 +/- 1.6 (placebo), P = 0.1]. Rates of lipid oxidation decreased [delta lipid oxidation (mg/kg x min), 0.02 +/- 0.14 (GH) vs. -0.32 +/- 0.13 (placebo), P <0.05], whereas glucose oxidation increased in the placebo-treated group (P <0.05). In the open phase, a decrease in insulin sensitivity was found in the former placebo group, although they lost body fat and increased fat-free mass [M-value (mg/kg x min), 5.1 +/- 0.7 (placebo) vs. 3.4 +/- 1.0 (open), P = 0.09]. In the group randomized to continued GH treatment almost all hormonal and metabolic parameters remained unchanged during the study. In conclusion, 1) discontinuation of GH therapy for 1 yr in adolescent patients induces fat accumulation without compromising insulin sensitivity; and 2) the beneficial effects of continued GH treatment on body composition in terms of decrease in fat mass and increase in fat-free mass does not fully balance the direct insulin antagonistic effects.