Objective: To study the reproducibility of 24-h serum growth hormone (GH) concentration profiles in adults.
Design: 24-h serum GH concentrations were constructed by drawing blood samples at 20-min intervals. Four study occasions over a period of 1 year were chosen to assess the reproducibility.
Subjects: Six healthy adult male volunteers of normal height and weight and aged between 20 and 22 years.
Measures: The resulting GH data arrays were analysed by Fourier transformation. Between and within individual variations were calculated and expressed in terms of coefficients of variation and data plotted to show variations between groups and individuals.
Results: The frequency component of GH secretion occurred with a dominant periodicity of between 160 and 240 min. Precise estimates of spectral power (strength of oscillatory activity) and period were obtained for group data, but such estimates cannot be inferred from a single profile. There was no significant difference in 24-h mean serum GH concentration over the year of study: occasion 1 (February) mean 2.0 mU/l (SD 0.5); occasion 2 (March) mean 3.7 mU/l (SD 2.8); occasion 3 (July) mean 2.7 mU/l (SD 1.4); occasion 4 (February) 2.5 mU/l (SD 1.5) (mean within individual coefficient of variation 35%, range 9-58). The concentrations of factors known to influence GH synthesis and secretion, insulin-like growth factor I, thyroxine, testosterone and oestradiol, varied little over the year of study.
Conclusions: These data demonstrate that group data are reproducible in terms of oscillatory activity and the amount of GH secreted and that 24-h GH profiles should be used predominantly for analysing group data. The variability between individual profiles limits their value in the investigation of children with growth failure and suspected GH insufficiency.