Do growth hormone (GH) serial sampling, insulin-like growth factor-I (IGF-I) or auxological measurements have an advantage over GH stimulation testing in predicting the linear growth response to GH therapy?

Summary objective To compare the relative utility of GH secretion via pharmacological stimulation, overnight serial sampling, IGF-I levels and auxological variables as predictors of change in height standard deviation score (ΔHt SDS) during GH treatment. design A multicentre observational study. patients Prepubertal children (n = 825) with idiopathic growth failure who were subsequently treated with GH were divided into two groups, based on their maximum GH response to pharmacological stimulation testing: (1) idiopathic GH deficiency (IGHD), defined by a maximum GH response < 10 µg/l (n = 300); and (2) idiopathic short stature (ISS), with a maximum GH response ≥ 10 µg/l (n = 525) (GH conversion factor: 3 IU = 1 mg). measurements Overnight spontaneous GH secretion was measured in all patients. The following characteristics of spontaneous GH secretion were studied: maximum or peak GH, mean peak GH, number of GH peaks, pooled GH, mean GH, and approximate entropy of GH secretion. results Although children with IGHD had lower indices of spontaneous GH secretion, there were no differences between IGHD and ISS groups in baseline Ht SDS, growth rate or IGF-I level. The dose and duration of GH therapy were similar. There was no statistically significant difference in the mean (± SD) change in Ht SDS (ΔHt SDS) in the two groups (IGHD 1·3 ± 0·9 and ISS 1·2 ± 0·8). Measures of spontaneous secretion, such as peak GH, mean of GH peaks, mean area under GH peaks, and mean GH, as well as IGF-I concentrations, were all statistically significantly correlated with ΔHt SDS in IGHD children (P < 0·0001). A significant correlation was also observed for pooled GH (P = 0·002) and approximate entropy (P = 0·01). Children with the most severe ISS (Ht SDS < −3·33) demonstrated a more disorganized pattern of GH secretion compared to children who were not as short (Ht SDS −2·33 to −1·64), as indicated by a higher approximate entropy (0·673 ± 0·193 vs. 0·607 ± 0·161, P < 0·004). This increased disorder in GH secretion was accompanied by lower IGF-I levels (104 ± 99 µg/l vs. 137 ± 74 µg/l, P < 0·001), even though pooled GH concentrations were indistinguishable between the two groups (2·2 ± 1·3 µg/l vs. 2·0 ± 1·0 µg/l). Children with IGHD demonstrated lower approximate entropy than did those with ISS (0·551 ± 0·235 vs. 0·631 ± 0·182, P < 0·0001). Duration of GH treatment, height deficit and genetic potential (midparental Ht SDS) were the most important variables influencing ΔHt SDS in children receiving GH therapy. Maximum stimulated GH, IGF-I and indices of spontaneous GH secretion also correlated with ΔHt SDS, but their relative importance varied among diagnostic groups. conclusions Patients with GH deficiency demonstrate a reduced capacity for GH secretion, while those with idiopathic short stature exhibit a more disorderly and less functional secretory pattern. Although effective in predicting a response to GH treatment in patients with severe GH deficiency, overnight serial sampling is less practical than other methods currently available. In addition, serial sampling was less useful as a predictor of growth response to exogenous GH in patients with idiopathic short stature.