This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ranke, M. B. Search for Related Content PubMed PubMed Citation Articles by Ranke, M. B. Related Collections Neuroendocrinology and Pituitary Pediatric Endocrinology

Increased Response, But Lower Responsiveness, to Growth Hormone (GH) in Very Young Children (Aged 0–3 Years) with Idiopathic GH Deficiency: Analysis of Data from KIGS

Pediatric Endocrinology Section, University Children’s Hospital, Eberhard Karls University (M.B.R.), D-72076 Tuebingen, Germany; Pfizer, Inc. (A.L., P.W.), S-11287 Stockholm, Sweden; Pediatric Growth Research Center, Department of Pediatrics, Queen Silvia Children’s Hospital (K.A.-W.), Sahlgrenska Academy of Goteborg University, S-416 85 Goteborg, Sweden; Department of Pediatrics, St. Mary’s Hospital (D.A.P.), Manchester M27 1HA, United Kingdom; and Baystate Medical Center Children’s Hospital, Tufts University of Medicine (E.O.R.), Springfield, Massachusetts 01106

Address all correspondence and requests for reprints to: Dr. Michael B. Ranke, Pediatric Endocrinology Section, University Children’s Hospital, Hoppe-Seyler Strasse 1, D-72076 Tuebingen, Germany. E-mail: michael.ranke{at}med.uni-tuebingen.de.

	Abstract

Top Abstract Introduction Subjects and Methods Results Discussion References

 
In children, GH secretion and sensitivity to GH are influenced by developmental changes. It is not clear whether the response to GH in very young children with GH deficiency (GHD) is the same as that in older, prepubertal children. A cohort of 265 children (180 males and 85 females) with idiopathic GHD from KIGS (Pfizer International Growth Database), with treatment started at less than 3 yr of age (mean age, 1.9 yr; group I) was compared with a cohort of 509 children (331 males and 178 females; group II) with treatment started at 7–8 yr of age (mean age, 7.5 yr). The following differences (P < 0.01) were found (given in mean values) between groups I and II at the start of GH treatment: 9% vs. 5% breech delivery, 38% vs. 14% multiple pituitary hormone deficiency, 4.2 vs. 5.9 ng/ml maximum GH in response to tests, –0.1 vs. –0.8 midparental height (MPH) SD score (SDS), –3.1 vs. –2.5 height SDS, 0.83 vs. 0.66 IU/kg·wk GH dose. After the first year of GH, the results were: 13.3 vs. 8.6 cm/yr height velocity, and 1.7 vs. 0.6 maximum change in height SDS. Using the previously developed growth prediction models for prepubertal children with idiopathic GHD more than 2 yr of age, our analysis revealed differences in the indexes of responsiveness in prediction models (Studentized residuals SDS, 0.7 vs.–0.3) and strikingly higher responsiveness to treatment among the young cohort, but with large scatter. Thus, new prediction models of height velocity (centimeters per year) were derived by means of multiple regression analysis for the young cohort, either involving (model A) or excluding (model B) the GH peak in tests. Model A explained 54% of the total variability with an error SD of 2.1 cm. Height velocity correlated with (parameters in order of importance) age (–), maximum GH (–), GH dose (+), weight SDS (+), height SDS minus MPH SDS (–), and birth weight SDS (+). Model B explained 45% of the total variability with an error SD of 2.3 cm. Height velocity correlated with (parameters in order of importance) age (–), GH dose (+), birth weight SDS (+), height SDS minus MPH SDS (–), and weight SDS (+). The predictors were qualitatively the same as those in the total prepubertal model involving all children more than 2 yr of age, but their quantitative impact in terms of partial contribution and the order of their importance were different for the young cohort. In particular, the partial contribution of the GH dose was higher, suggesting a greater gain in height per GH dose unit in the very young than in the older children. However, the rank order of the GH dose in the new models was lower, which suggests a slightly low sensitivity to GH in toddlers after the phase of severe GH insensitivity during early infancy. The early detection and GH treatment of congenital GHD is advantageous as a cost-effective strategy for achieving greater improvement of absolute height and growth velocity.

	Introduction

Top Abstract Introduction Subjects and Methods Results Discussion References

 
GH DEFICIENCY (GHD) in children is very heterogeneous in terms of its etiology, pathogenesis, and the age at which it is diagnosed. Even among children with evidence of congenital GHD, it is not unusual for the diagnosis to be established during a relatively late stage of childhood, at a time when short stature becomes obvious. There is, however, a small subset of patients whose GHD is recognizable during infancy or early childhood (i.e. <3 yr), and reports have shown that the early diagnosis of GHD and concomitant early initiation of treatment have proven effective in normalizing height (1, 2, 3, 4, 5). In attempting to individualize and optimize the outcome of GH replacement in GHD, prediction models have been derived from large cohorts of patients whose treatment began during childhood (1, 5, 6, 7). By means of this approach, it became clear that two treatment phases were crucial to the overall height outcome, namely, the prepubertal years and, in particular, the first year of treatment. Growth during the first years of life is characterized by a transition from the infancy to the childhood growth pattern and is probably governed by maturational processes within the GH-IGF system (8, 9, 10). Thus, it was not clear whether a growth prediction model devised for prepubertal children would be applicable to very young children with GHD. The aim of this study was primarily to examine whether the prediction models available for prepubertal children could be appropriately applied to very young children with GHD. In the event that these models proved unsuitable, we aimed to derive specific models for this age group. It was also the purpose of this analysis to gain additional insight into the potentially different mechanisms regulating the response to GH in early and later childhood.

	Subjects and Methods

Top Abstract Introduction Subjects and Methods Results Discussion References

 
We studied the observed and predicted growth response data (1) during the first year of GH treatment in patients with idiopathic GHD who were enrolled in KIGS (Pfizer International Growth Database) during the following periods: from birth to 3 yr of age and from 7–8 yr of age. The growth responses of patients in the group that started with very early GH replacement were correlated by means of multiple regression analysis, which included potentially relevant variables such as birth status, genetic background, treatment modality, and physical characteristics. The derivation of the growth prediction models was based on these variables in the very young group as well as in the prepubertal group in which GH treatment started after 2 yr of age (1).

Patients

We analyzed the data of all prepubertal patients with idiopathic GHD who had been enrolled in KIGS up to January 14, 2004. These patients had been treated exclusively with recombinant human GH (Genotropin, Pfizer, New York, NY) at six or seven injections per week. At this time the KIGS database included 265 children, aged 0–3 yr, and 509 children, aged 7–8 yr, who were suitable for analysis. The diagnosis had been made by the treating physician according to the KIGS etiology classification system (code 1) (11). A peak GH level in a provocation test of 10 µg/liter or less was an inclusion criterion for the analysis. All studied patients were in the prepubertal stage; i.e. boys had a mean testes volume of 3 ml or less, whereas girls had a Tanner breast stage of B1. For the calculation of a full year’s height velocity, height measurements were taken at intervals of 12–14 months. Patients were excluded from the analysis if they missed their GH injections for a total of more than 14 d in 1 yr. Patients born small for gestational age were excluded.

Statistical analysis

Growth responses (height velocities, centimeters per year) were correlated with several variables by means of multiple regression analysis. The mean ± SD of these variables are reported. SD scores (SDSs) were calculated as follows: SDS = (patient value – mean value for age- and sex-matched normal subjects) ÷ SD of the value for age- and sex-matched normal subjects. The variables tested were 1) status at birth: sex, weight SDS, length SDS, ponderal index, mode of delivery, and Apgar score; 2) genetic background: height SDS of the mother, height SDS of the father, midparental height (MPH) SDS, and ethnic origin (the ethnic background of the patients was analyzed by adding dummy variables, e.g. 0/1 Asian/not Asian, to allow mathematical analysis within the multiple regression computer program); 3) treatment modality: GH dose [international units (3 IU = 1 mg) per kilogram of body weight and international units per kilogram of ideal body weight (weight for height) per week] and frequency of GH injections; and 4) patient variables at start of treatment: age, height/length SDS, weight SDS, height/length SDS minus MPH SDS, peak GH level during provocative tests, and pituitary hormone deficiency status (i.e. isolated GHD or multiple pituitary hormone deficiencies).

The height/length standards used for normal children were those of Tanner et al. (12), and the weight standards were those of Freeman et al. (13). Birth weight for gestational age was transformed to SDSs based on the standards of Niklasson et al. (14). The MPH SDS was calculated as the father’s height SDS + mother’s height SDS = 1.61 (15), based on the standards of Tanner et al. (12).

The prediction models were developed by means of multiple linear regression analysis fitted by least squares and the REG procedure of the SAS computer program (mainframe version 6.12, SAS Institute, Inc., Cary, NC). A hierarchy of predictive factors was derived by the all possible regression approach, using Mallow’s C(p) criterion for ordering predictive factors, as described by Weisberg and Cook (16, 17). Differences between observed and predicted height velocities were expressed in terms of Studentized residuals. The residual is calculated as the observed height velocity minus the predicted height velocity for each observation, and the Studentized residual is the residual divided by its SE.

Wilcoxon rank tests were used for comparisons, median values, and 10–90th percentile range. Spearman correlation coefficients are quoted. The P values correspond to two-sided tests. In addition, means and SD values are given if appropriate. For multivariate regression analyses, the procedure REG in the program package SAS version 8 was used.

	Results

Top Abstract Introduction Subjects and Methods Results Discussion References

 
Demographic characteristics of the cohorts analyzed

The respective characteristics of the two groups of patients for the period preceding GH treatment are listed in Table la. Our findings showed that in both groups, gestational age, birth weight, birth length, and head circumference at birth were similar. In the younger cohort, the relative frequencies of breech delivery, hypoglycemia, microphallus (in males), and multiple pituitary hormone deficiency were significant. The relative frequency of cesarean sections, however, was not different between the two groups. In addition, the results of GH testing indicated that GHD in the younger cohort was more severe. In the older cohort, the heights of the mother and father and the MPH values were significantly lower.

View larger version (23K): [in this window] [in a new window]   FIG. 1. Height velocity during the first year of GH treatment in children with GHD less than 3 yr old and 7–8 yr old in relationship to normal references.

A significant difference in the Studentized residuals was found between the cohorts when the previously developed models were applied. In the younger group, the mean values were not only elevated compared with those in the older group (P < 0.001), but were also not similar to the previous model group due to the large scatter (Fig. 2). In comparison with the previous prediction models, developed on the basis of prepubertal children with ages above 2 yr, the predicted height velocities were slightly, but significantly (P < 0.05), higher than the observed height velocities (mean Studentized residual, –0.3; P < 0.01), but with a normal variability. This shows that the previously developed models are not applicable to the younger cohort of children less than 3 yr of age. We therefore developed specific prediction models for the group with patients younger than 3 yr of age (Fig. 2).

View larger version (32K): [in this window] [in a new window]   FIG. 2. Studentized residuals vs. predicted height velocity for the first year of GH treatment in 265 children with idiopathic GHD (0–3 yr of age) according to first year growth prediction models. A, Total prepubertal age model, including GH maximum; B, total prepubertal age model, excluding GH maximum (1 ); C, new model for the younger cohort, including GH maximum; D, new model for the younger cohort, excluding GH maximum.

The variables (mean and SD) found to be predictive of height velocity in our previous as well as new models for the first year of treatment, the overall correlation coefficients of the prediction models, and the error SD of their prediction as well as the probability level of differences in the various predictors used for comparison of the previous and present models are listed in Table 2. The order of importance in the models (rank) and the partially explained variability of the models (pR²) are also shown in Table 2.

When comparing the models that included the degree of GHD (GH peak in tests; model B; Table 2), the same predictors were also found to be significant (P < 0.001) in the two age groups. The model for the younger patients explained less (54% vs. 61%) of the total variability in the growth response, albeit with the same order of error. There were no differences in the quantitative effect of the distance to target height or of weight SDS. There was a lower effect from differences in weight at birth and differences in the degree of GHD (GH peak in tests), but a greater effect of differences in age (the younger the child, the higher the HT velocity) and differences in GH dose.

Thus, in both models, regardless of the GH peak in tests, the importance of the GH dose increased with respect to its rank in the model, but also in terms of its quantitative effect. During the second year of GH treatment, the data for the young cohort showed no difference between the observed height velocity and the height velocity predicted by the previously published model (data not shown here). Thus, the phenomenon of the difference between the very young and older, prepubertal children in terms of response to GH is restricted to the first year of treatment.

	Discussion

Top Abstract Introduction Subjects and Methods Results Discussion References

 
The ontogeny of the GH-IGF system during the postnatal months is characterized by elevated levels of GH secretion accompanied by impaired tissue GH receptor expression and low postnatal IGF-I levels; this is followed by a gradual decrease in GH secretion and an increase in IGF-I levels during infancy (10, 18). This adaptational process, which is assumed to be mainly the result of an increase in GH sensitivity, is mirrored by the pattern of growth during infancy, which is followed by another pattern representing the GH-sensitive period of later childhood (8). Although there are many studies of treatment of GHD children diagnosed during the midchildhood period, there is scarce literature about children with early-onset GHD and their responses to GH treatment (2, 3, 4, 5).

The cohort of children we analyzed in this study, who began GH treatment before 3 yr of age, shows the typical differences at birth and at the start of GH treatment compared with children diagnosed and treated during midchildhood. Size at birth was similar in the two cohorts, whereas parental height was lower in the group that started GH therapy later. In the early therapy group, the degree of GHD was more severe, and there was a higher fraction of patients with additional pituitary deficiencies; consequently, there was a higher prevalence of hypoglycemia and microphallus (in males) observed during infancy. The severity of the pituitary deficits in the young cohort probably resulted in a more profound and earlier stunting of growth. Birth history also revealed a higher proportion of breech deliveries in the young group, but the same frequency of cesarean section, although cesarean sections were more frequent in both GHD groups compared with a reference population (19). The potentially causal relationship between modalities of birth and lesions of the pituitary (20) remains obscure. In the group of children with early treatment, the applied dose of GH per kilogram of body weight was slightly higher than the dosage given during later childhood. This may partly be due to the fact that some KIGS investigators calculated the GH dose on the basis of body surface rather than weight.

To analyze the response to GH in a wide age range of prepubertal children by applying the previously developed growth prediction model, we chose a cohort of very young children (0–3 yr of age) and another cohort covering a narrow age slot (7–8 yr of age). Our findings showed that the mean values of the parameters characterizing the latter as well as those of the previously analyzed prepubertal cohort were very similar. Although there was no significant difference between the observed and the predicted height velocities during the first year of GH treatment in the 7–8 yr olds regardless of the model applied (either including or excluding GH peak in tests), there were significant differences in the younger cohort. The Studentized residuals, representing the index of responsiveness, were almost 1 SD above identity (zero), and the scatter was large (Fig. 2). It is remarkable that this was no longer the case during subsequent treatment, because a more uniform pattern of growth in response to GH therapy sets in after the first year of GH treatment. In view of the fact that the growth rate declines with age during prepubertal life, it becomes evident that the response to GH, in terms of centimeters per year or gain in height, is greater among younger patients. Height velocity, in terms of SDS, is lower in the younger cohort, a fact attributable to the greater SE of height velocity at that age.

On deriving new models for the young cohort, we did not anticipate that the same predictors were qualitatively relevant for the models of the young as well as for the total prepubertal cohort, nor did we expect that the overall predicted variabilities and errors of prediction would be very similar. However, the partial contributions of single predictors were found to be different. The fact that the influence of the differences in age was greater in the models is probably due to the normal pattern of growth during that age, which is characterized by a rapid decline in height velocity. The diminished effect of height distance to target height may be explained by the fact that the young organism seeks the growth target, which is an inherent process, which, however, does not end before the age of 3 yr.

In the model involving the peak GH levels in tests, these levels were of lesser significance for the young cohort. Although this parameter was found to be the most important during childhood, it dropped to the second rank and to only about one third of the partial contribution to the explained variability. Because the range in GH levels was the same in both groups, this finding may reflect the difficulty of establishing the quantitative level of GH deficiency in young children. Finally, the partial contribution to the explained variability in the growth response by the parameter of GH dose was found to be higher in the young cohort. This means that the gain in height per unit of GH given is higher in this age group. However, in both models, which either included or excluded the peak GH levels in provocation tests, the rank of importance of the GH dose within the models increased from rank 5 in the old models to ranks 2 and 3 in the new models, respectively. In the prediction models derived for prepubertal children with Turner syndrome (21) and short for gestational age (22), the GH dose is the most important predictor (rank 1) and indicates an impaired GH sensitivity if compared with GHD. This also suggests that at a mean age of 1.9 yr, the GH resistance during the growth phase of infancy has not yet completely subsided in the cohort. These observations do not contradict the view that GHD, as indeed all other growth disorders, should be discovered early by means of appropriately structured health programs. Early replacement of GH in younger children with GHD is more effective in achieving height improvement than when the treatment starts later in childhood. Apart from this, the psychological benefits are greater, and the improvement in cost-effectiveness is vast. The mechanisms that govern the responsiveness to GH in the young child need to be studied in large cohorts of affected children and should involve the standardized collection of additional anthropometrical and biochemical parameters.

	Acknowledgments

 
The authors are grateful to Priscilla Herrmann for her assistance in preparing this manuscript.

	Footnotes

 
First Published Online January 5, 2005

Abbreviations: , Maximum change; GHD, GH deficiency; MPH, midparental height; SDS, SD score.

Received June 3, 2004.

Accepted December 3, 2004.

	References

Top Abstract Introduction Subjects and Methods Results Discussion References

 

1. Ranke MB, Lindberg A, Chatelain P, Wilton P, Cutfield W, Albertsson-Wikland K, Price DA 1999 Derivation and validation of a mathematical model for predicting the response to exogenous recombinant human growth hormone (GH) in prepubertal children with idiopathic GH deficiency. KIGS International Board. Kabi Pharmacia International Growth Study. J Clin Endocrinol Metab 84:1174–1183[Abstract/Free Full Text]
2. Boersma B, Rikken B, Wit JM 1995 Catch-up growth in early treated patients with growth hormone deficiency. Dutch Growth Hormone Working Group. Arch Dis Child 72:427–431[Abstract/Free Full Text]
3. De Luca F, Arrigo T, Bozzola M, Crisafulli G, Ghizzoni L, Messina MF, Wasniewska M 1998 Hypopituitarism during the first year of life. A collaborative Italian study. Arch Pediatr 5(Suppl 4):313S–317S
4. Huet F, Carel JC, Nivelon JL, Chaussain JL 1999 Long-term results of GH therapy in GH-deficient children treated before 1 year of age. Eur J Endocrinol 140:29–34[Abstract]
5. Rappaport R, Mugnier E, Limoni C, Crosnier H, Czernichow P, Leger J, Limal JM, Rochiccioli P, Soskin S 1997 A 5-year prospective study of growth hormone (GH)-deficient children treated with GH before the age of 3 years. French Serono Study Group. J Clin Endocrinol Metab 82:452–456[Abstract/Free Full Text]
6. Ranke MB, Lindberg A, Guilbaud O 1994 Prediction of growth in response to treatment with growth hormone. In: Ranke MB, Gunnarsson R, eds. Progress in growth hormone therapy–5 years of KIGS. Mannheim, Germany: J & J Verlag; 97–111
7. Ranke MB, Lindberg A, Martin DD, Bakker B, Wilton P, Albertsson-Wikland K, Cowell CT, Price DA, Reiter EO 2003 The mathematical model for total pubertal growth in idiopathic growth hormone (GH) deficiency suggests a moderate role of GH dose. J Clin Endocrinol Metab 88:4748–4753[Abstract/Free Full Text]
8. Karlberg J, Albertsson-Wikland K 1988 Infancy growth pattern related to growth hormone deficiency. Acta Paediatr Scand 77:385–391[Medline]
9. Karlberg J 1989 A biologically-oriented mathematical model (ICP) for human growth. Acta Paediatr Scand 350(Suppl):70–94
10. Parks JS 2001 The ontogeny of growth hormone sensitivity. Horm Res 55(Suppl 2):27–31
11. Ranke MB 1991 The Kabi Pharmacia International Growth Study: aetiology classification list with comments. Acta Paediatr Scand 379(Suppl):87–92
12. Tanner JM, Whitehouse RH, Takaishi M 1966 Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children, 1965. II. Arch Dis Child 41:613–635
13. Freeman JV, Cole TJ, Chinn S, Jones PR, White EM, Preece MA 1995 Cross sectional stature and weight reference curves for the UK, 1990. Arch Dis Child 73:17–24[Abstract/Free Full Text]
14. Niklasson A, Ericson A, Fryer JG, Karlberg J, Lawrence C, Karlberg P 1991–1981 An update of the Swedish reference standards for weight, length and head circumference at birth for given gestational age 1997–1981. Acta Paediatr Scand 80:756–762
15. Ranke MB 1996 Towards a consensus on the definition of idiopathic short stature. Horm Res 45(Suppl 2):64–66
16. Weisberg S 1985 Applied linear regression, 2nd ed. Chichester, UK: Wiley & Sons
17. Cook RD, Weisberg S 1982 Residuals and influence in regression. New York: Chapman and Hall
18. Ogilvy-Stuart AL 2003 Growth hormone deficiency (GHD) from birth to 2 years of age: diagnostic specifics of GHD during the early phase of life. Horm Res 60(Suppl 1):2–9
19. Odlind V, Haglund B, Pakkanen M, Otterblad O 2003 Deliveries, mothers and newborn infants in Sweden, 1973–2000. Trends in obstetrics as reported to the Swedish Medical Birth Register. Acta Obstet Gynecol Scand 82:516–528[CrossRef][Medline]
20. Bierich JR 1992 Aetiology and pathogenesis of growth hormone deficiency. Baillieres Clin Endocrinol Metab 6:491–511[CrossRef]
21. Ranke MB, Lindberg A, Chatelain P, Wilton P, Cutfield W, Albertsson-Wikland K, Price DA 2000 Prediction of long-term response to recombinant human growth hormone in Turner syndrome: development and validation of mathematical models. KIGS International Board. Kabi International Growth Study. J Clin Endocrinol Metab 85:4212–4218[Abstract/Free Full Text]
22. Ranke MB, Lindberg A, Cowell CT, Albertsson-Wikland K, Reiter EO, Wilton P, Price DA 2003 Prediction of response to growth hormone treatment in short children born small for gestational age: analysis of data from KIGS (Pharmacia International Growth Database). J Clin Endocrinol Metab 88:125–131[Abstract/Free Full Text]

This article has been cited by other articles:

				L L Hui, C M Schooling, B J Cowling, S S L Leung, T H Lam, and G M Leung Are universal standards for optimal infant growth appropriate? Evidence from a Hong Kong Chinese birth cohort Arch. Dis. Child., July 1, 2008; 93(7): 561 - 565. [Abstract] [Full Text] [PDF]

				N. Bouhours-Nouet, F. Gatelais, F. Boux de Casson, S. Rouleau, and R. Coutant The Insulin-Like Growth Factor-I Response to Growth Hormone Is Increased in Prepubertal Children with Obesity and Tall Stature J. Clin. Endocrinol. Metab., February 1, 2007; 92(2): 629 - 635. [Abstract] [Full Text] [PDF]

				M. Bidlingmaier, J. Kim, C. Savoy, M. J. Kim, N. Ebrecht, S. de la Motte, and C. J. Strasburger Comparative Pharmacokinetics and Pharmacodynamics of a New Sustained-Release Growth Hormone (GH), LB03002, Versus Daily GH in Adults with GH Deficiency J. Clin. Endocrinol. Metab., August 1, 2006; 91(8): 2926 - 2930. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Submit a related Letter to the Editor Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Ranke, M. B. Search for Related Content PubMed PubMed Citation Articles by Ranke, M. B. Related Collections Neuroendocrinology and Pituitary Pediatric Endocrinology