| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Pediatrics (E.M.N.B., A.C.S.H.-K.), Division of Endocrinology, and Department of Epidemiology and Biostatistics (P.G.H.M.), Erasmus MC-Sophia Children’s Hospital, 3015 GJ Rotterdam, The Netherlands; and Department of Metabolic and Endocrine Diseases (J.v.D.), University Medical Center, Wilhelmina Children’s Hospital, 3508 AB Utrecht, The Netherlands
Address all correspondence and requests for reprints to: Anita C. S. Hokken-Koelega, M.D., Erasmus MC-Sophia Children’s Hospital, Department of Pediatrics, Division of Endocrinology, Dr. Molewaterplein 60, 3015 GJ Rotterdam, The Netherlands. E-mail: a.hokken{at}erasmusmc.nl.
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Objectives: The aim of the study was to investigate changes in free IGF-I during GH treatment in short SGA children and to evaluate whether free IGF-I levels contribute to predicting first-year growth response and/or adult height.
Design, Setting, and Intervention: We conducted a randomized, double-blind GH dose-response study with a GH dose of either 1 mg/m2·d (group A) or 2 mg/m2·d (group B). Free IGF-I, total IGF-I, and IGF binding protein (IGFBP)-3 were determined at baseline, after 1 and 5 yr, at stop, and 6 months after GH discontinuation.
Patients: We studied 73 (46 male) short SGA children (36 group A) with a baseline mean age of 7.7 (2.2) yr and a mean GH duration of 8.2 (2.1) yr.
Main Outcome Measures: Untreated SGA children had a mean free IGF-I SD score (SDS) of –0.2 (1.2), not related to total IGF-I. During GH therapy, free IGF-I significantly increased to 1.6 (0.7) SDS, as did total IGF-I and IGFBP-3 [2.0 (0.8) and 1.3 (0.9), respectively]. Multiple regression analysis showed that baseline free IGF-I and IGFBP-3 were negatively correlated with adult height SDS, whereas baseline bone age delay, target height SDS, baseline height SDS, and GH dose were positively correlated. Free IGF-I was also negatively correlated with first-year growth response.
Conclusions: Circulating baseline free IGF-I and IGFBP-3 were better predictors for adult height in GH-treated SGA children than total IGF-I, or total IGF-I to IGFBP-3 ratio. This suggests a possible role for free IGF-I measurement in predicting the effect of GH therapy in short SGA children.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
IGF-I plays an important role in both pre- and postnatal growth, and its serum levels are regulated by GH, and metabolic and genetic factors. In neonates born SGA, low-circulating IGF-I levels have been observed (7, 8, 9, 10). Short SGA children exhibit plasma IGF-I levels that are in the lower-normal range, approximately between 1 and 1.5 SD score (SDS) (11, 12, 13).
Several studies demonstrated that GH treatment in short SGA children results in normalization of height during childhood and adulthood (14, 15, 16, 17, 18). During GH treatment in SGA children, a rapid increase in total IGF-I levels and a slower increase in IGF binding protein (IGFBP)-3 levels occurs (19). Controversies exist on the role of total IGF-I and/or IGFBP-3 levels in the growth response. It has been reported that the growth response showed a positive association with the short-term increase in total IGF-I (19) and an inverse relation to baseline IGF-I levels (19, 20). Other authors (21) reported a positive correlation between the growth response and the change in both total IGF-I and IGFBP-3 in SGA children. However, in another study on short SGA children, one could not confirm these results (22). This apparent discrepancy may be due to the use of different definitions for SGA, leading to differences in inclusion criteria and study groups.
In the circulation, IGF-I is mainly bound to IGFBPs, of which six classes have been identified (IGFBP-1 to -6) (23). The GH-dependent IGFBP-3 is the major carrier protein of IGF-I in the circulation, normally accounting for more than 90% of the IGF binding (24). Under normal circumstances, less than 1% of the total plasma IGF-I pool is in the unbound free biologically active form (25, 26), which exchanges rapidly with the tissue compartments (23). To our knowledge, free IGF-I levels have not been studied in SGA children before and during GH treatment. The first aim of our study was to evaluate free IGF-I levels at start and during GH therapy until adult height. Second, we wanted to determine if free IGF-I levels would contribute to predicting the first-year growth response and adult height.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
A total of 90 short children born SGA participated in a multicenter GH trial in The Netherlands, which started in 1991. Inclusion criteria for the dose-response trial were: birth length SDS below –2; chronological age (CA) between 3 and 11 yr in boys and 3 and 9 yr in girls; height SDS for CA below –2; no spontaneous catch-up growth during last year; prepubertal stage; and uncomplicated neonatal period without severe asphyxia, sepsis, or lung problems (12, 15).
The GH trial evaluated the effect of GH on long-term growth and adult height. In brief, to stratify for spontaneous GH secretion during a 24-h GH profile, a group of 79 children was divided into three groups: "normal profile," "GH insufficient profile," and "no profile performed." Twenty-four-hour GH profiles were performed at baseline in 39 of the children, as previously described (27); an arginine stimulation test was performed in all 79 children. After stratification for age and spontaneous GH secretion during a 24-h GH profile, all children were randomly and blindly assigned to one of two GH-dosage groups: group A, 1 mg/m2·d (
33 µg/kg·d); or group B, 2 mg/m2·d (67 µg/kg·d) (15). GH was given double blindly. In addition, the remaining 11 short SGA children were treated parallel to the trial with a known dose of 2 mg/m2·d because they were older than the maximum age according to the inclusion criteria. Three monthly the total GH dose was adjusted to the calculated body surface. Biosynthetic GH (r-hGH Norditropin; Novo Nordisk A/S, Bagsvard, Denmark) was given sc once daily until adult height, defined as height velocity less than 0.5 cm during the previous 6 months. Every 3 months, height and weight were measured. Height was expressed as SDS for CA and gender, using the references for healthy Dutch children (28). Adult height was expressed as SDS for gender and an adult age of 20 yr.
Bone age (BA) was determined by the same investigator according to the Tanner and Whitehouse radius, ulna, short bones score.
Study subjects and design
For the present study, we only investigated patients who received GH therapy for 
5 yr. Seven of the 11 older subjects were excluded because they were treated less than 5 yr until adult height, whereas three other subjects were still being treated with GH at analysis and could, therefore, not be included. At the start of the present study, seven had dropped out of the GH trial and were lost to follow-up due to the following reasons: lack of motivation (n = 3), moving abroad (n = 2), treatment for precocious puberty (n = 1), and biochemical signs of GH insensitivity (n = 1). Thus, 73 subjects remained eligible for the present study.
Five of them were GH deficient (GHD), defined as having IGF-I and IGFBP-3 below –2 SD and peak GH levels less than 20 mU/liter after stimulation with arginine. Seven had a clinical diagnosis of Silver-Russell syndrome. All 73 subjects were prepubertal at start and after 1 yr of GH therapy.
Both EDTA plasma and serum samples were taken between 0900 and 1200 h at baseline, 1 and 5 yr after start of GH therapy, before discontinuation of GH therapy (final year), and 6 months after ending GH therapy (post GH). Thus, a maximum of five samples per subject were analyzed. The baseline, first-year, and post-GH samples were collected after an overnight fasting. All samples were stored appropriately at –80 C until assayed.
Levels of free IGF-I, total IGF-I, and IGFBP-3
Free IGF-I levels were measured in serum with a commercial two-site immunoradiometric assay (IRMA) using a commercial kit (Diagnostic System Laboratories, Inc., Webster, TX). This IRMA detects both the unbound IGF-I and the easily dissociable IGF-I (29). The interassay coefficient of variation was 9.7%, calculated from data produced by the investigators, measured at serum levels of 0.26 and 3.41 ng/ml. There was no difference in the coefficient of variation between the low and high level. To establish normative range values for circulating free IGF-I, serum samples were collected from healthy children (116 girls and 211 boys, aged between 0 and 17 yr) who underwent minor surgical procedures. Serum samples were obtained and stored at –80 C in well-capped tubes until analysis. Smoothed references for free IGF-I were constructed using the LMS method, designed for constructing normalized standards of nonparametric data, described by Cole (30).
The free IGF-I assays were performed by the same investigator (E.M.N.B.), in the same laboratory, under standardized circumstances, i.e. the assays were performed at 5 C, serum samples were kept at 5 C, and incubated for exactly 2 h. The various serum samples investigated in the present study had not been thawed previously. However, we did investigate the effect of repeated freezing and thawing (three times) of several (n = 13) serum samples on the free IGF-I levels but could not find a significant effect (i.e. not larger than the interassay variation). Moreover, samples tested again after a year of storage showed the same results. Using 13 healthy adult volunteers working in our department, we compared free IGF-I levels in sera obtained after an overnight fast with those in the nonfasting state. No significant differences were found. This finding is in agreement with a previous report on this subject, showing that, despite a marked elevation in IGFBP-1, overnight fasting did not influence circulating free IGF-I levels (31).
Total IGF-I plasma levels were measured by a semiautomated chemiluminescence technique (Advantage; Nichols Institute Diagnostics, San Juan Capistrano, CA). Plasma IGFBP-3 levels were measured by specific RIA, as described previously (32).
The LMS smoothed normative range values for total IGF-I, IGFBP-3, and the molar ratio between total IGF-I and IGFBP-3 have been previously determined in the same laboratory (32, 33).
Statistical methods
Unless indicated otherwise, results are expressed as mean (SD). Differences over time and between the GH-dosage groups were analyzed using repeated measurement analyses. Spearman’s 
correlations were used to assess correlations between SDSs of free IGF-I, total IGF-I, IGFBP-3, and total IGF-I to IGFBP3 molar ratio, and 24-h GH-profiles outcome measures. The outcome measures for the 24-h GH profiles were the number of peaks, mean pulse amplitude, and the area under the curve above zero level.
Multiple regression analyses were used to assess the relationships between either free IGF-I, total IGF-I, IGFBP-3, or total IGF-I to IGFBP3 ratio, and first-year change in height SDS and final height SDS outcome, respectively. Each laboratory parameter was entered separately in the model, along with age at start of GH therapy, target height (TH) SDS, and GH dose. In the model for adult height SDS, we also included baseline height SDS, and baseline BA delay [defined as CA minus BA (CA – BA)] was included, referring to the model described in a previous paper by our group (15). These multiple regression analyses were repeated for the subgroup of patients who underwent a 24-h GH profile. The 24-h GH profiles outcome measures were included into the models separately. The percentages of variance explained by the model (R2 in percentage) and the adjusted R2 (%) are given. A P value of <0.05 was considered statistically significant. All calculations were performed with SPSS 11.5 (SPSS, Inc., Chicago, IL).
Ethical considerations
The Medical Ethics Committee of each participating center approved the trial protocol. The Medical Ethics Committee of the Erasmus Medical Center approved the collection of reference serum samples in healthy children. Written informed consent was obtained from each participant older than 12 yr, and parents.
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
The clinical characteristics of the 73 children participating in this study are shown in Table 1
. There were no significant differences between the two GH-dosage groups. The difference in TH between the two GH-dosage groups nearly reached significance (P = 0.06).
|
The serum free IGF-I levels in the reference group showed an age-dependent increase (Fig. 1). Prepubertal girls had higher free IGF-I levels than boys, but during puberty, boys had higher free IGF-I levels than girls. At baseline, for both groups of short SGA children, serum levels of free IGF-I SDSs were not different from 0 SDS [mean SDS (SD): group A, –0.3 (1.3); group B, –0.2 (1.2)] (Fig. 2). There were also no significant differences between the two GH-dosage groups during GH treatment and thereafter. During GH therapy, the mean (SD) values of free IGF-I SDS in the total group of SGA subjects increased significantly from –0.2 (1.2) to 1.6 (0.7). After the first year, the free IGF-I SDS did not further increase. After 5 yr of GH treatment, 34.7% of the study subjects had free IGF-I SDS above the +2 SDS. Six months after the end of GH therapy, the mean free IGF-I SDS had decreased to 0.5 (1.0), which was significantly higher than baseline SDS. During GH therapy, 27.5% of all free IGF-I measurements were more than +2 SDS, 0.5% more than +3 SDS, and 32 of the 73 patients (44%) never exceeded the +2 SDS.
|
|
Total IGF-I and IGFBP-3 levels
At baseline, mean total IGF-I SDS and IGFBP-3 SDS were significantly lower than 0 SDS [–0.9 (0.9) and –1.2 (1.2), respectively], whereas mean SDS for the total IGF-I to IGFBP-3 molar ratio did not significantly differ from zero [–0.1 (0.2)] (Fig. 2
). In addition, for these parameters, no significant differences were noted between the two GH-dosage groups, at any point in time.
For the total group of SGA subjects, after 1 and 5 yr of GH therapy, mean total IGF-I SDS had increased to 1.4 (1.0) and 1.95 (0.8) SDS, mean IGFBP-3 SDS to 0.5 (1.1) and 1.2 (0.9), and the molar ratio SDS to 1.3 (1.6) and 1.3 (1.0), respectively.
Mean IGFBP-3 SDS after 5 yr of GH therapy of 1.2 (0.9) declined to 0.2 (1.0) SDS at GH discontinuation, which is in absolute values a decline from 139 (31) to 108 (40) nmol/liter. Mean IGFBP-3 SDS at GH discontinuation was significantly lower compared with the value found after 5 yr of GH therapy but did not significantly differ from that after 1 yr of GH therapy.
Six months after GH discontinuation, SDS for total IGF-I, IGFBP-3, and total IGF-I to IGFBP-3 molar ratio had decreased, but the values of these parameters were still higher than those found at baseline.
Relationships between free IGF-I, total IGF-I, IGFBP-3, and total IGF-I to IGFBP-3 molar ratio and height
Pretreatment. At baseline, there was a weak but significant correlation between free IGF-I SDS and height SDS (r = 0.25; P < 0.05). Baseline free IGF-I SDS was also positively correlated with baseline IGFBP-3 SDS (r = 0.41; P < 0.001), but not with baseline total IGF-I and the total IGF-I to IGFBP-3 molar ratio SDSs (r = 0.08, P = 0.49 and r = –0.17, P = 0.15, respectively). Baseline total IGF-I SDS, IGBP-3 SDS, and total IGF-I to IGFBP-3 molar ratio SDS did not show any correlation with baseline height SDS. At baseline, there were no correlations between any of the 24-h GH profile characteristics, or the stimulated GH peak and height SDS, free IGF-I, total IGF-I, IGFBP-3, or total IGF-I to IGFBP-3 ratio.
During GH treatment. After 1 and 5 yr of GH treatment, free IGF-I SDS showed a positive relationship with total IGF-I SDS [r = 0.46 (P < 0.001) and r = 0.32 (P < 0.01), respectively], and after 1 yr of GH with the SDS value of total IGF-I to IGFBP-3 molar ratio (r = 0.45; P < 0.001). There was no correlation between free IGF-I SDS and IGFBP-3 SDS after 1 and 5 yr of GH treatment.
First-year growth response.
For the change in height SDS during the first year, the multiple regression model, including the independent variables age at start of GH therapy, TH SDS, and GH dose, explained 38.8% (R2) of the variance (Table 2). Adding baseline free IGF-I SDS significantly increased the R2 to 48.3% (P < 0.01), contributing nearly 10% to the prediction of the first-year change in height SDS. When, instead of or in addition to free IGF-I, baseline total IGF-I SDS or IGFBP-3 SDS were added, this did not improve the model significantly (P = 0.4). First year’s change in free IGF-I SDS was not related to first-year change in height SDS. Adding baseline BA delay (CA – BA) did not improve the model.
|
Adult height SDS prediction.
For adult height SDS, the model, including the independent variables CA – BA at the start of GH therapy, TH SDS, height SDS at start, and GH dose, explained 41.9% (R2) of the variance in adult height SDS (adjusted R2: 38.3%). Adding baseline free IGF-I SDS, the R2 significantly increased to 49.5% (P < 0.01; adjusted R2 45.5%). Subsequent addition of baseline IGFBP-3 SDS improved the model even further (P = 0.01; R2 55.9%; adjusted R2 51.6%). The addition of baseline total IGF-I SDS did not improve the model further (P = 0.5). The optimal multiple regression models for first-year response and adult height are shown in Table 2.
When free IGF-I was not included in the model, baseline total IGF-I together with the independent variables in a multiple regression model also improved the model significantly, but less pronounced (maximum R2 was 45%).
None of the 24-h GH-profile characteristics were of significant influence on adult height outcome.
|
Discussion |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Baseline free IGF-I was inversely related to the first-year growth response and adult height in contrast to baseline levels of total IGF-I. In addition, IGFBP-3 SDS was inversely related to adult height SDS. All of the characteristics of the 24-h GH profiles were inversely related to first-year growth response, but not to adult height or any of the baseline IGF-I parameters.
For the present study, we used, under strictly controlled conditions, the Diagnostic System Laboratories, Inc., IRMA for the measurement of free IGF-I levels in serum. In fact, this assay determines the sum of truly free IGF-I and readily IGFBP-dissociable IGF-I, which is thought to represent a more biologically relevant pool than free IGF-I alone (25). There was no difference between fasting and nonfasting state. Serum-free IGF-I levels in healthy boys and girls appeared to increase with age, and were higher in prepubertal girls than in boys, which is in agreement with earlier reports (25, 26, 35).
Most samples were analyzed at the end of the study, several years after collecting them. Total IGF-I and IGFBP-3 in EDTA plasma or serum is proved to be very stable after long-term storage at –80 C. We had the opportunity to test free IGF-I after up to 3 yr of storage; the samples showed no significant differences. The tested samples had not previously been thawed and were stored in well-capped tubes at –80 C.
At baseline mean, serum free IGF-I SDS in short SGA children was not significantly different from 0 SDS, either with or without GHD. However, serum levels of total IGF-I and IGFBP-3 were significantly lower than normal, which has previously been reported (13, 19, 22, 36). In our study there were also no differences in free IGF-I levels between SGA children with and without GHD during GH therapy and thereafter. Because free IGF-I levels are decreased in untreated GHD children and adults (25, 35, 37), our results suggest that short SGA children with GHD have a different pathophysiology than children with idiopathic GHD.
Baseline free IGF-I SDS correlated inversely with first-year growth response and adult height SDS, also after adjusting for baseline levels of total IGF-I and IGFBP-3. This suggests that at baseline, the free fraction of the total IGF-I level in the circulation has a predictive value not only for first-year growth response to GH therapy, but also for adult height SDS in short SGA children, regardless of baseline total IGF-I levels. IGFBP-3 was of significant influence on adult height, not on first-year growth response.
We did not find a correlation between baseline free IGF-I SDS and baseline total IGF-I SDS. This finding is in contrast with observations under conditions of primary abnormalities in the GH secretion, such as GHD and acromegaly, in which usually a close relationship is found between serum free IGF-I levels and total IGF-I levels and total IGF-I to IGFBP-3 molar ratios (38). One explanation might be that short stature in SGA is due to genetic variations in the IGF-I gene, resulting in altered IGF-I binding to their various binding proteins (39, 40).
There was no correlation between the change in free IGF-I and first-year growth response. Apparently, the relationship between the concentration of free IGF-I in the circulation and growth (tissue) response is different before and during GH treatment. Because total IGF-I increases more than free IGF-I during GH treatment, it might be that there is an increased clearance of free IGF-I from the circulation to the tissues, due to a very short half-life of free IGF-I in the circulation compared with that of total IGF-I (38). However, the nearly equal SDS levels of total IGF-I and free IGF-I during GH treatment do not support this theory.
However, one has to consider that the levels of IGF-I or IGFBP-3 measured were circulating levels. Unfortunately, we could not measure locally produced IGF-I levels. In this respect, it must be emphasized that circulating concentrations of IGF-I (and IGFBP-3) may not necessarily reflect those at the various tissue compartments because there is also a (unknown) contribution of locally produced growth factors.
In the present model, including the variables baseline BA delay, TH SDS, baseline height SDS, and GH dose, the unstandardized coefficient (B) of the GH dose was 0.40 (SE, 0.188). The coefficient increased to 0.55 after adding baseline free IGF-I and IGFBP-3 to the model. The GH dose now had a significant influence on adult height, whereas in a previous paper, the GH dose did not reach significance (P = 0.2) (15). This difference might be explained by the larger study population in the present paper (n = 77). Our data are in line with results from a metaanalysis (n = 82) of data of several SGA studies by de Zegher and Hokken-Koelega (41), who found that the higher GH dose showed a difference in height gain of 0.4 SDS compared with the lower GH dose. However, the metaanalysis also indicated that height gain was less GH-dose dependent over the long-term than over the short-term.
This is the first report on free IGF-I levels in GH-treated SGA children. Notably, free IGF-I levels were identical for both GH-dosage groups. Data on a relationship between free IGF-I levels and long-term cancer risk are lacking. For that reason, it is not yet possible to draw definite conclusions about the usefulness of free-IGF-I measurements as a marker of long-term safety of GH treatment. Therefore, we cannot conclude from our results that the higher GH dose is as safe as the lower GH dose.
In conclusion, our study showed that untreated short children born SGA have normal free IGF-I levels, whereas total IGF-I and IGFBP-3 were decreased. Baseline free IGF-I and IGFBP-3 were better predictors for adult height in GH-treated SGA children than were total IGF-I or total IGF-I to IGFBP-3 ratio. Free IGF-I was also a predictor for short-term growth. This finding suggests a role for free IGF-I measurement in the prediction of growth response of short SGA children at the start of GH therapy.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online May 15, 2007
Abbreviations: BA, Bone age; CA, chronological age; GHD, GH deficient; IGFBP, IGF binding protein; IRMA, immunoradiometric assay; SDS, SD score; SGA, small for gestational age; TH, target height.
Received March 29, 2006.
Accepted May 3, 2007.
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. Maiorana and S. Cianfarani Impact of Growth Hormone Therapy on Adult Height of Children Born Small for Gestational Age Pediatrics, September 1, 2009; 124(3): e519 - e531. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
, Significantly different from baseline. #, Significantly different from group A. (If GH discontinuation was at 5 yr, patient was only included in "at stop" sample, not in both "after 5 yr" and "at stop".)