Appraisal of growth hormone (GH) secretion: evaluation of a composite pharmacokinetic model that discriminates...

Criteria for a diagnosis of GH deficiency include inadequate GH secretion as assessed by provocative testing. The changes in serum GH concentration in such tests, however, do not uniformly predict treatment responses. We questioned whether the changes in serum GH have a uniform dependence among subjects on the mass of secreted GH. We simulated spontaneous GH secretory events with bolus infusions of recombinant human GH (rhGH) in 15 somatostatin-infused adult subjects. Maximum serum GH responses, the GH areas under the curve, and GH mass calculated from deconvolution techniques are all indexes of GH secretion influenced by GH clearance and distribution volume. In this group, these indexes showed a nonuniform dependence on the known GH dose. Despite somatostatin infusion, we found evidence for low level basal GH secretion with oscillatory characteristics that may have influenced the GH concentration dependence on GH dose. We then developed and evaluated a new pharmacokinetic model to account for pulsatile, basal, and oscillatory inputs to the serum GH concentration profile. The new model is comprised of three terms. The first describes plasma GH concentrations from exogenous administration of rhGH according to a one- or a two-compartment model. The second term accounts for basal GH secretion. The third is a cosinor function that describes the oscillatory pattern of basal GH. The composite pharmacokinetic model predicted plasma GH concentrations well (r2 = 0.88-0.97); pharmacokinetic and cosinor parameters had high precision and narrow 95% confidence intervals. The pharmacokinetic parameters were stable and independent. The mean values and coefficients of variance (SD/mean) of GH pharmacokinetic parameters in our 15 subjects were: clearance, 0.236 L/min (24%); volume of distribution, 3.46 L (30%); and terminal half-life, 12.3 min (37%). The values for the cosinor parameters were: basal concentration, 0.22 ng/mL (85%); amplitude, 0.758 (50%); cycle, 121 min (27%); and time shift (acrophase), 60.3 min (53.6%). During the 9-h study, clearance decreased from 0.259 +/- 0.09 to 0.214 +/- 0.06 L/min (P < 0.03), and basal concentration increased from 0.20 +/- 0.22 to 0.33 +/- 0.33 ng/mL (P < 0.5). We conclude that our model can provide useful estimates of GH pharmacokinetics in the presence of basal, oscillating, endogenous concentrations without administering a dose of radiolabeled GH. The substantial inter- and intrasubject variance in pharmacokinetic parameters between subjects negates the assumption of a uniform relationship between GH secretion and serum GH concentration and detracts from the utility of a GH concentration cut-off point in GH testing. These findings have implications to the valid appraisal of GH deficiency states, selection of rhGH treatment candidates, and physiological regulation of the GH axis.