Although GH replacement improves the features of GH deficiency (GHD) in adults, it has yet to be established whether cessation of GH at completion of childhood growth results in adverse consequences for the adolescent with GHD. Effects of continuation or cessation of GH on body composition, insulin sensitivity, and lipid levels were studied in 24 adolescents (13 males, 11 females, aged 17.0 +/- 0.3, yr, mean +/- se, puberty stage 4 or 5) in whom height velocity was less than 2 cm/yr. Provocative testing confirmed severe GHD [peak GH < 9 mU/liter (3 microg/liter)] in all cases and was followed by a lead-in period of 3 months during which the pediatric dose of GH continued unchanged. Baseline investigations were then performed using dual-energy x-ray absorptiometry (body composition), lipid measurements, and assessment of insulin sensitivity by both homeostasis model assessment and a short insulin tolerance test. Twelve patients remained on GH (0.35 U/kg.wk), and 12 patients ceased GH treatment. The groups were followed up in parallel with repeat observations made after 6 and 12 months. No endocrine differences were evident between the groups at baseline. GH cessation resulted in a reduction of serum IGF-I Z score [-1.62 +/- 0.29, baseline vs. -2.52 +/- 0.12, 6 months (P < 0.05) vs. -2.52 +/- 0.10, 12 months (P < 0.01)] but values remained unchanged in those continuing GH replacement. Lean body mass increased by 2.5 +/- 0.5 kg ( approximately 6%) over 12 months in those receiving GH but was unchanged after GH discontinuation. Cessation of GH resulted in increased insulin sensitivity [short insulin tolerance test, 153 +/- 22 micromol/liter.min, baseline vs. 187 +/- 20, 6 months (P < 0.05) vs. 204 +/- 14, 12 months (P = 0.05)], but no significant change was seen during 12 months of GH continuation. Lipid levels remained unaltered in both groups. Continuation of GH at completion of linear growth resulted in ongoing accrual of lean body mass (LBM), whereas skeletal muscle mass remained static after GH cessation in these adolescents with GHD. This divergence of gain in LBM is of potential importance because increases in LBM occur as a feature of healthy late adolescent development. GH is a major mediator of insulin sensitivity, independent of body composition in adolescents. Further studies are required to determine whether discontinuation of GH in the adolescent with severe GHD once linear growth is complete results in long-term irreversible adverse physical and metabolic consequences and to determine conclusively the benefits of continuing GH therapy.