Abstract
Growth hormone (GH) secretion rates were estimated from 24-hour plasma GH concentration profiles using a model-based numerical deconvolution method. The basic kinetic parameters describing GH distribution and elimination were obtained from studies using single intravenous injections in GH-deficient children. Computer simulation techniques were used to produce 24-hour GH secretion profiles, which could serve as reference curves for analysis of the impact of certain practical aspects of the estimation of GH secretion, such as the method of blood sampling (discrete or integrated), sampling frequency and analytical errors. The results show that sampling intervals of 20 min are acceptable even in the presence of analytical errors, and that with this sampling interval, discrete blood sampling does not seem to be preferable to continuous withdrawal of blood.