Mean plasma GH concentrations are controlled by the frequency, amplitude, and duration of underlying GH secretory bursts as well as by the half-life of endogenous GH. We investigated the specific mechanisms that subserve the clinically recognized negative effects of age and adiposity on mean serum GH concentrations. To this end, 21 healthy men, aged 21-71 yr, who were of nearly normal body weight underwent blood sampling at 10-min intervals for 24 h. Deconvolution analysis was used to estimate specific features of GH secretion and clearance. Compared to younger men, the older tertile of men had significant reductions in 1) GH secretory burst frequency, 2) the half-life of endogenous GH, and 3) the daily GH secretory rate, but not 4) GH secretory burst half-duration, amplitude, or mass. Linear regression analysis disclosed that age was a major negative statistical determinant of GH secretory burst frequency (r = -0.80; P = 0.005) and endogenous GH half-life (r = -0.70; P = 0.024). Body mass index, an indicator of relative obesity, was a significant negative correlate of GH half-life (P = 0.045) and GH secretory burst amplitude (P = 0.031). Age and body mass index each correlated negatively with the daily GH secretion rate (P = 0.0031 and P = 0.027, respectively), and together accounted for more than 60% of the variability in 24-h GH production rates (r = -0.78; P = 0.00056). On the average, for a normal body mass index, each decade of increasing age attenuated the GH production rate by 14% and the GH half-life by 6%. Conversely, each unit increase in body mass index, at a given age, reduced the daily GH secretion rate by 6%. We conclude that age and relative adiposity are distinct and specific correlates of individual attributes of GH secretion and clearance in men.

We have examined the mechanisms underlying reduced circulating GH concentrations in the obese human. Computer-assisted (deconvolution) analysis was used to determine endogenous GH secretory and clearance rates quantitatively from entire 24-h plasma GH concentration profiles. These analyses revealed that the half-life (t 1/2) of endogenous GH was significantly shorter in obese (11.7 +/- 1.6 min) than in normal weight subjects (15.5 +/- 0.81 min; P less than 0.01). The accelerated blood disposal rate of GH was not due to decreased circulating concentrations of GH-binding protein, since the latter were similar in obese (25 +/- 1.0%) and normal weight (24 +/- 2.3%) men. However, obese men had significantly fewer GH secretory bursts (3.2 +/- 0.53 vs. 9.7 +/- 0.67/day; P less than 0.01). Among the rare GH secretory bursts that occurred in obese subjects, there were significantly prolonged mean intersecretory burst intervals (282 +/- 65 vs. 131 +/- 11 min; P less than 0.05). The resultant daily GH production rate in obese men was reduced to one fourth that in normal weight individuals. Both GH secretion rate and burst frequency were negatively correlated with the degree of obesity (ponderal index). The decreases in GH burst frequency and half-life were specific, since GH secretory pulse amplitude (maximal rate of GH release), the mass of GH released per burst, and the duration of computer-resolved GH secretory bursts were not different in obese and normal weight men. We conclude that obese men harbor a double defect in GH dynamics involving both GH secretion and clearance, and that the severity of the GH secretory deficiency is proportionate to the degree of obesity.

We have examined the mechanism subserving the in vivo circadian rhythm of cortisol in men. To this end, blood samples were withdrawn at 10-min intervals for 24 h in each of six men to yield well-defined profiles of episodic cortisol release. A novel multiple-parameter deconvolution model was applied to discriminate the number, amplitudes, and durations of all statistically significant underlying cortisol secretory bursts from the plasma hormone concentrations and simultaneously estimate the endogenous half-life of cortisol disappearance in each subject. These experiments disclosed randomly occurring cortisol secretory bursts at a mean frequency of 19 +/- 0.82 events per day (interpulse interval 77 +/- 4.0 min). Secretory bursts exhibited a mean half-duration (duration at half-maximal amplitude) of only 16 +/- 0.61 min indicating that 95% of daily cortisol secretion occurred in 8.2 h. Cortisol secretory burst frequency varied 2.2-fold over 24 h, whereas cortisol secretory burst amplitude varied 6.6-fold. We conclude that the nyctohemeral pattern of cortisol variation in vivo can be accounted for by an amplitude-modulated, random burstlike mode of cortisol secretion without the need to postulate a tonic mode of cortisol release.