OBJECTIVE To estimate the half-life of growth hormone in young adult patients with type I (insulin-dependent) diabetes mellitus following bolus injection and prolonged exposure for the purpose of deconvolution analysis of plasma growth hormone profiles to determine growth hormone secretory rates. METHODS In the bolus study, an intravenous bolus injection of 100 mU of biosynthetic human growth hormone was given while endogenous growth hormone was suppressed by a continuous infusion of somatostatin under three different glucose clamp conditions: normoglycaemia (5 mmol/l) with normoinsulinaemia (65 pmol/l); hyperglycaemia (12 mmol/l) with normoinsulinaemia; and normoglycaemia with hyperinsulinaemia (360 pmol/l). In the infusion study, the effect of prolonged and repeated growth hormone exposure upon the growth hormone half-life was estimated. Three pulses of 60 minutes growth hormone infusion (6 mU/kg/pulse) two hours apart under euglycaemic somatostatin suppression were applied. METHODS Six young adult patients with type I (insulin-dependent) diabetes mellitus were studied in both the bolus and the infusion study. RESULTS Mean GH half-lives by mono-exponential analysis were not significantly different remaining unaltered by the short-term metabolic changes of hyperglycaemia and hyperinsulinaemia. Data were therefore pooled yielding an overall mean GH half-life of 13.6 minutes (range 11.9-19.4). Applying a bi-exponential model mean GH half-lives were 3.1 minutes (range 2.5-5.9) for the rapid phase of distribution of the hormone and 13.8 minutes (range 9.6-16.9) for the decay of GH from the circulation. The GH half-life during the infusions studies did not vary with repeated exposure but was significantly longer (mean half-life of 25.7 minutes; range 19.4-37.1) than during the bolus studies (P less than 0.001). CONCLUSIONS The half-life of exogenous r-hGH is not affected by glucose or insulin concentrations but increases after prolonged GH exposure in young adults with type I (insulin-dependent) diabetes mellitus.