Serum aminoterminal propeptide of type III procollagen (PIIINP) levels, a reliable marker of collagen formation, were evaluated in children (C=7) and adults with childhood-onset (CO=10) and acquired (A=18) GH deficiency (GHD) before, during and after withdrawal of rec-GH therapy (C=0.6 IU/kg/week, CO=0.5 IU/kg/week, A=0.25 IU/kg/week). The duration of treatment was 12 months for C and A and 6 months for CO; investigations were carried out before and at 3, 6, 9 and 12 months (for C and A) and at 3 and 6 months (for CO) of GH treatment and 6 months after the withdrawal of therapy (for A and CO). Data obtained from patients were compared with those recorded in two age- and sex-matched control groups. Before treatment, serum PIIINP levels were significantly lower (p<0.001) in C with GHD (mean+/-SE: 2.9+/-0.4 ng/ml) than in controls (6.1+/-0.4 ng/ml), while no significant differences were recorded between adults with CO/A-GHD (3.7+/-0.5 ng/ml and 3.4+/-0.2 ng/ml) and controls (3.2+/-0.2 ng/ml). GH treatment caused a significant increase (p<0.0001) of PIIINP levels both in C (3rd month: 4.4+/-0.2 ng/ml, 6th month: 5.1+/-0.4 ng/ml, 12th month: 5.1+/-0.5 ng/ml), CO-GHD (3rd month: 12.7+/-1.2 ng/ml; 6th month: 10.2+/-0.6 ng/ml) and A-GHD (3rd month: 10.0+/-1.0 ng/ml; 6th month: 8.4+/-0.6 ng/ml; 12th month: 7.0+/-0.7 ng/ml), the increase being dose-dependent (more marked and sustained in adults with CO-GHD). The maximal stimulation of collagen synthesis occurred after 3 months of GH treatment in adults with GHD, while a more gradual and less relevant increase was observed in C with GHD. Six months after the withdrawal of GH therapy, serum PIIINP levels of adults with CO-GHD (3.6+/-0.3 ng/ml) were similar to those recorded before treatment, while in adults with A-GHD serum PIIINP levels (2.6+/-0.2 ng/ml) were significantly lower (p<0.01) than in basal condition. In conclusion, our study shows that: a) GHD is associated with a reduction of soft tissue formation in children, while it seems to exert no relevant effects in adults with GHD; b) GH therapy causes a rapid stimulation of collagen turnover, which shows a different pattern in children and adults; c) the GH-induced stimulation of collagen synthesis is rapidly removed after the withdrawal of GH treatment. For these reasons, the determination of peripheral markers of GH effects appears useful for the monitoring of GH therapy and can contribute to assess the "tailored" substitutive dose for the individual patient.