Growth hormone normalises height, prediction of final height and hand length in children with Prader-Willi syndrome after 4 years of therapy

BACKGROUND: Based on the reported favourable effects of growth hormone (GH) treatment on growth and body composition in Prader-Labhart-Willi syndrome, we studied age dependency and the long-term effects on growth dynamics to elucidate the assumed hypothalamic GH deficiency. METHODS: We examined 23 children treated with hGH (24 U/m(2)/week) during a median of 4 (range 1.5-5.5) years; group 1: 10 young underweight (age 0.3-4.1 years), group 2: 8 prepubertal overweight (age 3.7-9.5 years) and group 3: 5 pubertal overweight children (age 9.0-14.6 years). RESULTS: After 4 years of therapy, height gain amounted to 1.8 SD; height (0.0 SD) and hand length (-0.2 SD) were normalised in the 2 prepubertal groups; in children above 6 years, height prediction approached parental target height. Weight for height rose in group 1 (to 0.64 SD) and decreased in group 2 (to 0.71 SD) to normal levels. Bone maturation of the pubertal children was too advanced to show a clear growth response to GH (height gain 0.42 SD). Even in this group, weight for height was reduced, but remained supernormal. CONCLUSION: Under exogenous GH, growth and body proportions are normalised in prepubertal children. With early institution of treatment, final height prediction reaches the parental target height range after 3 years. Such a growth-promoting effect of exogenous GH has so far only been described in children with GH deficiency.     

Growth failure in early life: an important manifestation of Turner syndrome

The goals of this study were to test the hypothesis that girls with Turner syndrome (TS) experience growth failure early in life and to establish model-based normative growth charts for 0- to 8-year-old American girls with TS. Full-term girls with TS who had 5 or more measurements of height obtained during their first 10 years of life prior to initiation of growth hormone, estrogen and/or androgen therapy were eligible for this study. A nonlinear mixed-effects model comprising the first two components of the infancy-childhood-puberty (ICP) model of growth was fitted to the longitudinal height measurements and compared with those of healthy American girls. Height measurements (n = 1,146) from 112 girls with TS (45,X: 57.1%; 45,X/46,XX: 12.5%; 46,X, iso(X): 4.5%, and other: 25.9%) were analyzed. Mean height SDS fell from -0.68 at birth to -1.60 at 1 year, -1.80 at 2 years and -1.95 at 3 years. When compared to controls (676 girls, 4,537 measurements), girls with TS grew more slowly due to three principal factors: a slow growth rate of the infancy component, a slow growth rate at the onset of the childhood component, and delayed onset of the childhood component. Traditional concepts of growth failure in TS should be revised. Physicians should consider the diagnosis of TS in any girl with unexplained failure to thrive or short stature, even in the first 3 years of life.     

Serum IGF-I and IGFBP-3 levels of Turkish children during childhood and adolescence: establishment of reference ranges with emphasis on puberty

AIMS/METHODS: We established age- and sex-related reference ranges for serum insulin-like growth factor-I (IGF-I) and insulin-like growth factor binding protein-3 (IGFBP-3) levels in 807 healthy Turkish children (428 boys, 379 girls), and constructed a model for calculation of standard deviation scores of IGF-I and IGFBP-3 according to age, sex and pubertal stage. RESULTS: Serum IGF-I and IGFBP-3 concentrations tended to be higher in girls compared to boys of the same ages, but the differences were statistically significant only in pubertal ages (9-14 years) for IGF-I and only in prepubertal ages for IGFBP-3 (6-8 years) (p < 0.05). Peak IGF-I concentrations were observed earlier in girls than boys (14 vs. 15 years, Tanner stage IV vs. V) starting to decline thereafter. IGFBP-3 levels peaked at age 13 and at Tanner stage IV in both sexes with a subsequent fall. Serum levels of IGF-I and IGFBP-3 increased steadily with age in the prepubertal stage followed by a rapid increase in IGF-I in the early pubertal stages. A relatively steeper increase in IGF-I but not in IGFBP-3 levels was observed at age 10-11 years in girls and at 12-13 years in boys which preceded the reported age of pubertal growth spurt. At late pubertal stages, both IGF-I and IGFBP-3 either did not change or decreased by increasing age. Interrelationships between growth factors and anthropometric measurements have been described, and the physiologic consequences of these have been discussed in detail. CONCLUSIONS: Differences in the pattern of IGF-I and IGFBP-3 in the present paper and those reported in other studies emphasize the importance of locally established reference ranges. Establishment of this reference data and a standard deviation score prediction model based on age, sex and puberty will enhance the diagnostic power and utility of IGF-I and IGFBP-3 in evaluating growth disorders in our population.     

Pubertal growth spurt induced by human chorionic gonadotropin in hypogonadotropic growth hormone-deficient children

Eight hypogonadotropic growth hormone-deficient children were treated with human chorionic gonadotropin (HCG) while they continued to receive a fixed dose of HGH for a one year period. They were observed for changes in somatomedin C (IGF-I) and height increase velocity. Mean somatomedin C was 0.79 +/- 0.30 U/ml in normal prepubertal children (N = 7) and 0.78 +/- 0.31 U/ml in prepubertal normal short children (N = 22). At pubertal stage 3, somatomedin C was 2.21 +/- 1.23 and 2.05 +/- 0.44 U/ml in normals (N = 5) and in normal short children (N = 7), respectively. When 3000-5000 units/week of HCG were given to each of the 8 hypogonadotropic growth hormone-deficient children who were receiving HGH at a mean dose of 0.33 +/- 0.05 IU/kg/week, testosterone increased from less than 0.3 ng/ml to more than 5 ng/ml at 6 months in 3 cases and at 12 months in 2 cases, while the testosterone concentration was less than 3.5 ng/ml in the remaining 3 cases. The rate of height increase rose significantly (p less than 0.001) from 5.2 +/- 1.0 to 9.3 +/- 1.4 cm/year mimicking the normal pubertal growth spurt. However, the mean somatomedin C concentration was 0.44 +/- 0.23 before therapy, 0.33 +/- 0.30 at 6 months and 0.31 +/- 0.14 U/ml at 12 months after the start of HCG therapy. It is concluded that the pubertal growth spurt induced by HCG in hypogonodotropic GH-deficient male children is not mediated by the increase in somatomedin C production.     

Effects of sport (football) on growth: auxological, anthropometric and hormonal aspects

There are very few data available on the relationship between sporting activities, endocrine levels and changes in anthropometric measurements during growth. In order to study these relationships, we have made measurements of growth, changes in physical conformation and the plasma levels of several hormones [cortisol, dehydroepiandrosterone sulphate (DHEA-S), testosterone, growth hormone, somatomedin C, insulin, glycaemia and haemoglobin A1C] in 175 boys, aged 10-16 years, who have played football at a competitive level and in 224 boys, severing as controls, who have never performed sporting activities regularly. The football players were divided into prepubertal and pubertal subjects (10-11.99 years, 12-13.99 years and 14-16 years, chronological and bone age groups). Our results showed no significant differences in the growth indices between prepubertal athletes and controls, but the plasma level of DHEA-S was significantly higher (P less than 0.05) in the athletes. Pubertal football players, however, were significantly taller than the control subjects, particularly at 14-16 years chronological age. There were no such significant differences when bone age was considered. The pubertal football players were also more advanced in all biological indices of maturity, i.e. pubic hair, testicular volume and bone age. The increase in DHEA-S in pubertal football players, already seen in prepubertals, was also combined with a significant increase in testosterone, growth hormone and cortisol levels. Thus, in football players the DHEA-S level is already higher during prepuberty. This increase thus precedes all other indices of growth and maturation associated with puberty. We hypothesize that, while not excluding the possible influence of selection, as ours is a cross-sectional study, adrenal hyperactivity may be mainly responsible for the earlier onset of pubertal growth and maturity in exercising males.     

Growth curves of children with Down syndrome.

Growth curves of 105 children with Down syndrome (50 boys and 55 girls) were established. At birth height, weight and head circumference of Down syndrome children were lower than these parameters in controls. This delay remained stable until puberty. For weight there was no clear-cut pubertal growth spurt. For stature, the prepubertal growth spurt occurred earlier (at the age of 11 years in boys and 9 1/2 years in girls) than in controls but was less marked. As a result, Down syndrome patients had a short stature with a quite normal weight. These reference curves, available since prenatal diagnosis of Down syndrome is performed routinely, are helpful for monitoring normal and abnormal development in Down syndrome patients.     

Physiological growth hormone secretion in children with short stature and intra-uterine growth retardation

31 prepubertal children with short stature [mean height standard deviation score (SDS) -2.84] and low birth weight (mean -2.82 SDS) were studied. Mean age was 6.0 years and mean height velocity SDS was -0.76. Patients were classified as having either the clinical characteristics of Russell-Silver syndrome (RSS) (4 F, 13 M) or not (4 F, 10 M). All children had an overnight profile of spontaneous growth hormone (GH) secretion. 4 children achieved a maximum GH concentration of less than 20 mU/l. 9 children with RSS secreted only one large GH peak during the night. Most of the non-RSS group had normal GH pulse frequency but 3 boys had a fast-frequency pattern. Abnormal GH secretion may contribute towards growth failure in children with low birth weight/RSS.     

Assessment of prolactin secretion in children: a profile of circadian prolactin secretion and the principles for interpreting it

INTRODUCTION: Prolactin (Prl) is secreted in a circadian pattern, although no method of interpreting it has yet been established. The aim of the study was to assess Prl secretion in children on the basis of the Prl circadian profile and to establish principles for the interpretation of the results obtained by this method. MATERIAL AND METHODS: The analysis comprised 41 healthy short children (25 boys); aged 5.2-16.3 years, in whom hormonal disorders and chronic diseases had been excluded. The children were divided into prepubertal or pubertal subgroups. Serum Prl concentrations were measured every 3 hours for 24 hours. To assess the rhythm the parameters of macroscopic analysis were calculated and receiver operating characteristic (ROC) analysis was performed. The group for comparison consisted of 30 children aged 8.9-17.2 years with hyperprolactinaemia. RESULTS: In each subgroup significantly higher Prl concentrations were observed at night than by day. No statistical differences were noticed between the groups regarding Prl concentrations at particular time points or parameter values during circadian Prl rhythm evaluation. In the group analysed weak correlations were found between age and Prl peak and trough levels. On the basis of ROC analysis criteria for the existence of normal circadian Prl rhythm in children were established. CONCLUSIONS: 1. The presence of normal circadian Prl rhythm is observed if at least one of the following three criteria is fulfilled: amplitude >1.8779; X(n)/X(d) ratio >1.685; regression index <-0.4107. 2. No interpretation in relation to sex, age and stage of puberty is necessary for the circadian prolactin secretion rhythm in children.     

Plasma levels of insulin-like binding protein-2 in prepubertal short children and its diagnostic value in the evaluation of growth hormone deficiency

AIM: This study was designed to investigate whether determination of plasma insulin-like growth factor (IGF)-binding protein-2 (IGFBP-2) levels could be of benefit in the evaluation of childhood growth hormone (GH) deficiency (GHD). METHOD: A retrospective analysis was performed on 91 prepubertal children referred for investigation of short stature. Maximal GH levels in plasma after provocative stimuli were between 1.0 and 93.0 mU/l, 6 subjects exhibiting peak values of <5 mU/l. Initially a GH peak of 20 mU/l was used as a cutoff limit to define GHD and idiopathic short stature (ISS) patients. The results of GH provocative tests were compared to age- and gender-based standard deviation scores (SDS) of plasma IGFBP-2, IGF-I, IGFBP-3 and the molar ratios of the latter two to IGFBP-2. The respective normative range values for these parameters were determined in plasma samples from 353 healthy children (i.e. 171 girls, 182 boys). RESULTS: Circulating IGFBP-2 levels did not correlate with height SDS, height velocity SDS or the peak GH levels after provocative stimuli. A weak negative relationship was found between IGFBP-2 and IGF-I. Plasma levels of IGFBP-2 in GHD patients were higher than those of ISS children, who had normal levels. Although at the optimal cutoff point of -0.71 SDS 91.5% of the GHD patients were identified correctly, a substantial proportion (71.9%) of the ISS subjects also had IGFBP-2 levels above this limit. The use of various combinations of IGFBP-2, IGF-I, IGFBP-3 and the derived ratios only slightly improved the diagnostic efficiency as compared to the results of the individual tests. Neither IGFBP-2 nor the IGFBP-3/IGFBP-2 and IGF-I/IGFBP-2 ratios were found to be related to the short- (1 year) or long-term (3 years) growth response to GH therapy. CONCLUSION: It is concluded that none of the tests investigated, either alone or in various combinations, are reliable in either predicting the peak GH level after provocative stimuli in prepubertal short children or in predicting their growth response to GH.     

Effect of chronic clonidine treatment on urinary growth hormone excretion and linear growth in children with short stature

Eleven prepubertal children with short stature were treated with clonidine (0.15 mg/m2 daily) for a period of 1 year. The effect of this drug was evaluated on both clinical (growth velocity, height standard deviation scores for chronological age and bone age) and hormonal (urinary growth hormone excretion and insulin-like growth factor I) parameters. Our study shows that long-term clonidine administration in children with short stature did not result in significant differences in growth velocity, height standard deviation scores for chronological age and bone age, insulin-like growth factor I or in urinary growth hormone excretion.     

Differences in infantile growth patterns in Turner syndrome girls with and without spontaneous puberty.

OBJECTIVE: The role of prepubertal estrogen in child growth was modeled using Turner's syndrome, comparing growth patterns of girls who later did or did not enter puberty spontaneously. The hypothesis was that TS patients with normal prepubertal estrogen levels would have a different growth pattern from those with subnormal estrogen levels. STUDY DESIGN: Growth data from 78 full-term patients with Turner's syndrome were collected retrospectively. 24/78 later developed spontaneous puberty, (+Pub), and their growth data were compared to TS patients without spontaneous puberty (-Pub). A nonlinear mixed model was fitted using the bi-exponential model. RESULTS: The growth velocity difference between the -Pub and +Pub groups suggests an early infantile growth advantage in the -Pub group, which disappears before the end of the first year of life; growth velocity remains similar (+/- 1 cm/y) for the next 6 years and declines at age 7-8 years in the +Pub group faster than it does in the -Pub group. Bi-exponential analysis showed that both the 1st (restrictive) and 2nd exponent (forward) were different (p = 0.0003). CONCLUSIONS: Comparison of girls with or without spontaneous puberty suggests a role for estrogen in child growth. Estrogens restrict infantile growth, as well as growth during the mid-childhood spurt.     

Validity of height velocity as a diagnostic criterion for idiopathic growth hormone deficiency and Turner syndrome.

A deflecting growth curve over several years is sometimes the only indication for the possible presence of a growth disorder. In this study we looked at the potential diagnostic role of long-term downward deflection of the growth curve. It reports on the diagnostic validity of height velocity over 1, 2 or 3 years for isolated idiopathic growth hormone deficiency and for Turner syndrome in prepubertal children with a height that is still above -2.5 standard deviation scores (SDS). 1-year height velocity was found to have no diagnostic value because of an almost complete overlap of height velocity distributions with normal prepubertal children. However, height velocity over 3 years was found to have an acceptable validity in children 5-12 years old. In this age range a change in height SDS of -0.75 can be used as a valid criterion for further examination of karyotype and GH secretion capacity even if there are no other clear signs of a particular growth disorder.     

Sequelae of syndrome X in children born small for gestational age

OBJECTIVE: Low birth weight is associated with the presence of syndrome X in adults. We studied the components of this syndrome in prepubertal children born SGA (small for gestational age) and children born AGA (appropriate for gestational age). METHODS: Twenty-nine SGA children, age (mean +/- SD) 9.1 +/- 1.1 years and 24 AGA children, age 9.0 +/- 1.1 years were studied. Fasting serum lipid concentrations were determined. A hyperinsulinemic euglycemic clamp was performed to measure insulin sensitivity. Ambulatory monitoring was performed to obtain 24-hour recordings of blood pressure. RESULTS: Prepubertal SGA children are less insulin sensitive and have a higher nighttime systolic blood pressure (SBP) after correction for BMI than children born AGA. No differences were found in lipid concentrations between the 2 groups. CONCLUSIONS: Not all components of syndrome X can yet be found in 9-year-old children born SGA; follow-up of this cohort is required.     

Effects of growth hormone therapy on glucose metabolism and insulin sensitivity indices in prepubertal children with Prader-Willi syndrome

In Prader-Willi syndrome (PWS) growth hormone therapy (GHT) improves height, body composition, agility and muscular strength. In such patients it is necessary to consider the potential diabetogenic effect of GHT, since they tend to develop type 2 diabetes, particularly after the pubertal age. The aim of our study was to investigate the effects of GHT on glucose and insulin homeostasis in PWS children. An oral glucose tolerance test (OGTT) was performed in 24 prepubertal PWS children (15 male, 9 female, age: 5.8 +/- 2.8 years), 16 were obese (group A) and 8 had normal weight (group B), before and after 2.7 +/- 1.3 years GHT (0.22 +/- 0.03 mg/kg/week) and, only at baseline, in 35 prepubertal children with simple obesity (19 male, 16 female) (group C). Fasting glucose and insulin, glucose tolerance, insulin sensitivity index (ISI), homeostasis model assessment of insulin resistance (HOMA-IR), quick insulin check index (QUICKI), area under the curves (AUC) of glucose and insulin were estimated. At the start of GHT, all PWS children were normoglycaemic and normotolerant but two developed impaired glucose tolerance after 2.2 and 1.9 years of therapy, respectively. At baseline, group A showed lower fasting insulin levels, HOMA-IR and AUC of insulin, higher ISI, QUICKI and AUC of glucose than group C. Comparing groups A and B, AUC of insulin was higher and ISI lower in group A. During GHT, a significant increase of fasting insulin and glucose, a worsening of insulin resistance (HOMA-IR) and insulin sensitivity (QUICKI) was found only in group A while ISI did not change. The AUC of glucose decreased in both groups instead AUC of insulin did not change. BMI-SDS decreased in group A and increased in group B. The increased insulin resistance and decreased insulin sensitivity in obese PWS patients, as well as the occurrence of impaired glucose tolerance during GHT, suggest that a close monitoring of glucose and insulin homeostasis is mandatory, especially in treated obese PWS children.     

Progressive deceleration in growth as an early sign of delayed puberty in boys

OBJECTIVE: To describe the prepubertal growth pattern in boys with delayed puberty. METHODS: Growth curves for height and height velocity covering the age range 4-14 years were constructed on the basis of retrospectively obtained data in 85 boys with delayed puberty, who attained a normal final height. RESULTS: Between the age of 4 and 14 years the height in this cohort progressively deviated from the normal reference. At the age of 4 years, the height SDS was already significantly lower (median -0.8; p < 0.001) and progressively diminished during childhood, resulting in a median height SDS of -1.1 at the age of 12 years (p < 0.001). The median final height of this cohort (-0.4) was not different from their target height (-0.2). The degree of deceleration in growth during childhood was not determined by birth weight or birth height and did not influence final height. The decline of the height velocity with age in this group of boys with delayed puberty was significantly smaller (p < 0.001) than predicted by the model of Rikken and Wit. CONCLUSION: Late-maturing boys often show a prepubertal deceleration in growth that starts at an early age but that does not affect final height.     

Serum leptin in short children born small for gestational age: dose-dependent effect of growth hormone treatment

OBJECTIVE: To study the effects of different regimens of growth hormone (GH) treatment on serum leptin levels in 78 short prepubertal children born small for gestational age (SGA). METHODS: The children were originally included in two independent multicenter trials, one in Belgium and one in the Nordic countries. SGA children were randomized either to remain untreated or to be treated with GH at a daily dose of 0.1, 0.2 or 0.3 IU/kg for 2 years. Thereafter, treatment was continued for another 2 years in the Nordic children, whereas it was discontinued in the Belgian children. RESULTS: In the GH treatment groups, a significant dose-dependent decrease in leptin levels was found during the first year of therapy, with a mean decrease of 13, 23 and 32% in the groups receiving GH at 0.1, 0.2 and 0.3 IU/kg, respectively. When high-dose treatment was interrupted, serum leptin increased within 1 year to pretreatment levels. CONCLUSION: Serum leptin levels in short children born SGA are transiently reduced by GH treatment in a dose-dependent fashion. The most pronounced changes in serum leptin were documented within the first year after initiation and withdrawal of high-dose GH treatment.     

Growth and somatotype patterns of manus children,territory of Papua and New Guinea: Application of a modified somatotype method to the study of growth patterns

The growth and somatotype patterns of Manus children, Territory of Papua and New Guinea, reflected in anthropometric measurements and somatotype photographs, is reported for 438 children, ranging from 15 months of age to maturity. Somatotype distributions and individual somatotype photographs indicate little difference between boys and girls from age 1 to 4 years; between age 5 and 9 years boys shift toward higher second component ratings; between age 10 and 18 years boys shift toward higher second and third component ratings, while girls shift toward higher first and third component ratings. Four sample somatotype photographs are representative of the population at age 10 and 14 years, and show age and sex characters easily noted visually. Comparison of height, weight and subscapula skinfold values show that the Manus children resemble the Kaiapit and Bundi children of the New Guinea highlands populations reported by Malcolm in the form of their growth curves but not in growth rates. Manus children grow more slowly than the British, but faster than the Kaiapit and Bundi children. While low skinfold values and height/weight ratios derived from means for height and weight suggest somatotypic similarity of the Manus, Kaiapit and Bundi populations, somatotype photographs are needed for confirmation. Lacking historical information and genetic evidence, we can speculare that diet, nutrition, climate and perhaps physical activity are factors which influence the growth patterns of these populations. The protein intake derived from fish and other seafood may be an important factor in the relatively fast growth and earlier maturation of the Manus. Malcolm thinks that low protein intake plays an important role with the Kaiapit and Bundi, but does not rule out survival value of slow growth and short stature.     

Association of pharmacological tests and study of 24-hour growth hormone secretion in the investigation of growth retardation in children: analysis of 257 cases

Growth hormone (GH) secretion can presently be investigated by several methods: pharmacological provocative tests, study of 24-h GH secretion, measurement of somatomedin-C (Sm-C)/insulin-like growth factor (IGF) I, and the growth hormone-releasing hormone (GHRH) test. In order to compare the results obtained, these methods were used in 257 children with growth retardation (169 boys, 88 girls). Their height SD was -2.7 +/- 0.2, chronological age 11 3/12 +/- 1 6/12 years, and bone age 8 4/12 +/- 1 4/12 years. Mean growth velocity was 4.5 +/- 1.5 cm/year. One hundred and thirty-eight boys and 80 girls were prepubertal, and 31 boys and 8 girls were pubertal (B2 G2). All children underwent the study of 24-h GH secretion (n = 257) and pharmacological provocative tests (two tests, n = 213; one test n = 44). Sm-C/IGF I was measured in prepubertal children (n = 131), and a GHRH test was carried out (n = 153). In addition, the mean integrated concentration of growth hormone secretion (IC-GH) was assessed in a control group of 23 children and was found to be 5.4 +/- 1.2 ng/ml/min. The IC-GH in the group as a whole was 2.6 ng/ml/min. The mean maximum peak during pharmacological tests varied considerably according to the test used, ranging from 7.8 ng/ml for the arginine test to 17.1 ng/ml for the glucagon and betaxolol test. The maximum peak and the 24-h IC-GH were not significantly correlated.(ABSTRACT TRUNCATED AT 250 WORDS)     

Age at Puberty and the Emerging Obesity Epidemic

Background

Recent studies have shown that puberty starts at younger ages than previously. It has been hypothesized that the increasing prevalence of childhood obesity is contributing to this trend. The purpose of this study was to analyze the association between prepubertal body mass index (BMI) and pubertal timing, as assessed by age at onset of pubertal growth spurt (OGS) and at peak height velocity (PHV), and the secular trend of pubertal timing given the prepubertal BMI.

Methodology/Principal Findings

Annual measurements of height and weight were available in all children born from 1930 to 1969 who attended primary school in the Copenhagen municipality; 156,835 children fulfilled the criteria for determining age at OGS and PHV. The effect of prepubertal BMI at age seven on these markers of pubertal development within and between birth cohorts was analyzed. BMI at seven years was significantly inversely associated with age at OGS and PHV. Dividing the children into five levels of prepubertal BMI, we found a similar secular trend toward earlier maturation in all BMI groups.

Conclusion/Significance

The heavier both boys and girls were at age seven, the earlier they entered puberty. Irrespective of level of BMI at age seven, there was a downward trend in the age at attaining puberty in both boys and girls, which suggests that the obesity epidemic is not solely responsible for the trend.     

Effect of growth hormone on short normal children.

The growth of 26 short normal prepubertal children (mean age 8.4, height velocity standard deviation score for chronological age between +0.4 and -0.8) was studied for two years. Sixteen children were treated with somatrem (methionyl growth hormone) during the second year, and the remaining 10 children served as controls. During one year of treatment the height velocity standard deviation score for chronological age increased from the pretreatment mean of -0.44 (SD 0.33) to +2.20 (1.03). These values represented a change in height velocity from a pretreatment mean of 5.3 cm/year (range 4.6-6.9) to 7.4 cm/year (range 5.7-9.9). In the control group the height velocity standard deviation score was unchanged. Bone age advanced by 0.75 (0.33) years in the treated group compared with 0.70 (0.18) years in the control group. There was a significant increase in the height standard deviation score for bone age (0.63 (0.55] in the treated group. Multiple regression analysis of predictive factors contributing to the change in height velocity standard deviation score over the first year of treatment showed that the dose of growth hormone and pretreatment height velocity standard deviation score were important, together yielding a regression correlation coefficient of 0.80. The only metabolic side effect of treatment was an increase in fasting insulin concentration, which may be an important mediator of the anabolic effects of growth hormone. Treatment had no effect on thyroid function, blood pressure, or glucose tolerance. At the end of the treatment year seven of the 16 treated children had developed antibodies to growth hormone, but they were present in low titre with low binding capacity and in no child was growth attenuated. Biosynthetic growth hormone improved the height velocity of children growing along or parallel to the third height centile, but the effects on height prognosis need to be assessed over a longer period.