P-093. non-parametric deconvolution provides an objective assessment of gh responses to gh-releasing stimuli in normals

The applicability of a new non-parametric deconvolution (DEC) to the analysis of GH 'response-to-stimuli' data was studied. Thirty healthy adults (I: n = 30) were tested with a single 1 ?g/kg/bw GHRH bolus; group II (n = 6) was tested, on separate days, with 2 consecutive 1 ?g/kg/bw GHRH (a) and hexarelin (b) boluses. A numerically efficient regularization-based non-parametric DEC algorithm incorporating non-negativity constraints was used to estimate the time profile of the instantaneous secretion rate (ISR) and its integral. Results: the analysis of the benchmark problem showed that the proposed method is capable of providing an accurate reconstruction of the ISR. The analysis of the data showed a clear rise in the ISR after the first bolus, with most of the response within 60? from the stimulus. In group I, discarding the samples collected at 90-120? only marginally affected the estimate of the cumulated 0?-60? ISR (variation always <3%). The analysis of repeated stimuli showed that GH secreted after the 2nd boluses was clearly reduced vs that after the 1st ones. In conclusion, even with relatively few samples, non-parametric DEC is capable of providing a reliable, smooth, and non-negative estimate of the GH-ISR that offers a valid representation of the GH secretory dynamics.