Objective. Although short children who were born small for gestational age (SGA) seem to have normal body proportions, objective data both before and during growth hormone (GH) treatment are very limited. Therefore, we investigated in a large group of short children who were born SGA the effects of GH treatment versus no treatment on head circumference (HC) and body proportions. Furthermore, we studied differences in linear growth and HC between SGA children who were born with a low birth length and birth weight (SGAL+W) and SGA children who were born with a low birth length only (SGAL).
Methods. An open-labeled, GH-controlled, multicenter study was conducted for 3 years. Non–GH-deficient short SGA children (n = 87), with a mean age (standard deviation) of 5.9 (1.5) years, were randomized to either a GH group (n = 61), receiving GH in a dose of 33 μg/kg/day, or an untreated control group (n = 26). Height; weight; HC; sitting height; armspan; and hand, tibial, and foot size were measured and expressed as standard deviation score (SDS) adjusting for gender and age.
Results. At baseline, all anthropometric measurements, except HC SDS, were significantly lower compared with −2 SDS. During GH treatment, all anthropometric measurements normalized in accordance to the normalization of height SDS. At the start of the study, mean HC SDS was significantly lower in SGAL+W children compared with SGAL children. It is interesting that most (14 of 16) children with an HC SDS less than −2.00 had been born SGAL+W. During GH treatment, the 3-year increase in height, HC, and other anthropometric measurements was comparable between SGAL+W and SGAL children. In both SGAL+W and SGAL control subjects, no changes in SDSs of height, HC, and other anthropometric measurements were found during the 3-year follow-up period.
Conclusions. Untreated short SGA children have normal body proportions with the exception of HC, which is relatively large in many of these children. SGAL+W children still had a smaller HC at the age of 5.9 years compared with SGAL children. Three years of GH treatment induced a proportionate growth resulting in a normalization of height and other anthropometric measurements, including HC, in contrast to untreated SGA control subjects.
Short stature occurs in 10% to 15% of children who were born small for gestational age (SGA).1,2 The pathophysiology underlying this insufficient catch-up growth is still unknown, but disturbances in the growth hormone (GH)/insulin-like growth factor-I (IGF-I) axis may play a role.3,4
Several studies have shown beneficial effects of GH treatment on height in these short children.5–7 GH treatment resulted in a significant increase in height to values within the normal range followed by growth along the target height (TH) percentile.
Although short children who were born SGA seem to have normal body proportions, objective data are limited. Sas et al8 reported in a group of short SGA children a normalization of body proportions during GH treatment with either a dose of 1 or 2 mg/m2 body surface area/day. Because this study did not comprise an untreated control group, a causal effect of GH treatment on the normalization of body proportions could not be proved. There are no data on longitudinal changes of body proportions in short, untreated SGA children.
In most GH trials in short SGA children, SGA is defined by either birth length and birth weight alone or both. Obtaining an accurate length at birth is more difficult than obtaining an accurate weight at birth. However, when performed by trained individuals, birth length measurements are reliable. None of the GH trials has investigated differential effects of GH on growth and hormonal changes in SGA children who were either born with both a low birth length and a low birth weight (SGAL+W) or born with a low birth length only (SGAL). It is generally thought that in those who were born SGAL+W, growth restriction occurred early in gestation, whereas in SGAL newborns, this occurred later in gestation. It has been reported that during childhood, SGA children have, in general, a smaller head circumference (HC) compared with children who are appropriate for gestational age.9,10 No data are available on HC in short SGA children before and during GH treatment.
We present the results of a randomized, 3-year, controlled GH trial that studied HC and body proportions in short SGA children at baseline and during 3 years of GH treatment in comparison with results of untreated short SGA children. Furthermore, we studied linear growth and HC in SGAL+W versus SGAL children. Our study is the first to differentiate between SGA children who had been proportionately (SGAL+W) or disproportionately (SGAL) small at birth.
The study comprised 91 Dutch children (43 boys and 48 girls) who had short stature and were born SGA. All children fulfilled the same inclusion criteria: 1) birth length standard deviation score (SDS) less than −1.88 (ie, below third percentile) for gestational age11; 2) an uncomplicated neonatal period, without signs of severe asphyxia (which was defined in the present study as Apgar score <3 after 5 minutes), sepsis, or long-term complications of respiratory ventilation such as bronchopulmonary dysplasia; 3) chronological age between 3.00 and 7.99 years at start of the study; 4) height SDS for age less than −1.88 according to Dutch standards12; 5) height velocity SDS for age <0 to exclude children with spontaneous catch-up growth; 6) prepubertal, defined as Tanner stage 1 or a testicular volume <413; 7) sufficient GH response to a GH-stimulation test with either clonidine or arginine, which was defined as a GH peak >10 μg/L; and 8) normal liver, kidney, and thyroid functions. Children with endocrine or metabolic disorders, chromosomal defects (eg, Turner syndrome), and growth failure caused by other syndromes (eg, emotional deprivation, severe chronic illness, chondrodysplasia), with the exception of Silver-Russell syndrome, were excluded.
The study was approved by the Ethics Committees of all 9 participating centers. Written informed consent was obtained from the parents or custodians of each child.
The study design is an open-labeled, multicenter study with a randomized control group. All children (n = 91) were stratified according to age (3.00–5.50 vs 5.50–7.99) and height of the parents (height of both parents more than −1.88 SDS vs height of at least 1 parent less than −1.88 SDS). After stratification, the patients were randomly assigned to either the GH group (two thirds of children) or the control group (one third of children). The GH group (n = 64; 27 boys and 37 girls) started immediately with GH treatment at a dose of 1 mg/m2 body surface area/day (∼33 μg/kg/day). The control group (n = 27; 16 boys and 11 girls) remained untreated for 3 years and subsequently received the same GH treatment as the GH group. Biosynthetic GH (r-hGH Norditropin; Novo Nordisk A/S, Copenhagen, Denmark) was given subcutaneously once daily at bedtime. Once per 3 months, the GH dose was adjusted to the calculated body surface area.
Standing height was measured once per 3 months by 2 trained investigators (N.A., later V.B.) using a Harpenden stadiometer, and values were expressed as SDS for gender and chronological age (height SDS) using Dutch references.12 TH was calculated using Dutch reference data according to the formula: 1/2 * (Hfather [in cm] + Hmother [in cm] + 13) + 4.5 cm for boys and 1/2 * (Hfather [in cm] + Hmother [in cm] − 13) + 4.5 cm for girls, where the addition of 4.5 cm represents the secular trend. TH was expressed as SDS using Dutch references.12 Weight for height (WH) was used as a measure of body composition because body mass index is not an accurate estimate of body composition during GH treatment in short SGA children (Arends et al, submitted). WH was expressed as SDS for gender and age.12 HC was measured once per 3 months and expressed as SDS for gender and age.12
Sitting height (SH) was measured every 6 months using a Harpenden SH table. Every 6 months, armspan, the length of the left hand (hand), left foot (foot), and left tibia (tibia) were measured by the same investigators using a Harpenden anthropometer. All measurements were expressed as SDS adjusting for gender and age. Reference data were available from the Dutch Oosterwolde study, which consisted of 1240 healthy boys and 1093 healthy girls.14
Definition of SGAL+W and SGAL
SGAL+W was defined as a birth length and a birth weight −2.00 SDS or lower for gestational age.1,11 SGAL was defined as a birth length −2.00 or lower and a birth weight more than −2.00 SDS for gestational age.
Of 91 children, 4 dropped out of the study for the following reasons: 1 child was very disappointed that she was randomized into the control group, 2 children had psychological problems with the daily GH injections, and in 1 child celiac disease was diagnosed. Because these children dropped out either at start or during the first year of the study, these children were excluded from baseline and 3-year analysis. Therefore, 87 children were included for statistical analysis.
During the 3-year study period, puberty started in 4 children of the GH group and in 2 children of the control group. Analysis was performed in prepubertal children only. As soon as a child entered puberty, he or she was excluded from additional analysis. Onset of puberty was defined as a breast development stage 2 according to Tanner scale for girls and a testicular volume ≥4 mL for boys, determined by means of a Prader orchidometer.
Data were expressed as the mean ± SD. SDSs were compared with 0 using 1-sample t test. Differences between groups were tested using independent t tests. Differences in 1-year changes between the groups were tested using analysis of covariance, and differences between points in time within the groups were tested by paired t tests. Spearman correlation coefficient was used for correlations. Statistical significance was defined as P < .05. Statistical tests were performed with use of SPSS package (version 10.0).
Table 1 shows the baseline characteristics of the total, the GH group, and the control group. At baseline, no significant differences were found between the 2 groups. Thirteen children had Silver-Russell syndrome.
At baseline, no significant differences in anthropometric measurements were found between the 2 groups (Table 2). All anthropometric measurements, except HC SDS, were significantly lower compared with −2 SDS. Remarkably, HC SDS was the least affected part of the body and was not significantly lower than −2.00 SDS. Sixteen (18%) of 87 children had an HC SDS of −2.00 or lower. Height SDS, SH SDS, and foot SDS were most severely reduced. Height SDS correlated significantly with SH (r = 0.6; P < .001), foot (r = 0.6; P < .001), tibia (r = 0.7; P < .001) and armspan (r = 0.5; P < .001). However, height SDS correlated neither with hand nor with HC SDS.
During GH treatment, the size of the measured body parts expressed as SDS showed an increment toward 0 (Fig 1). After 2 years of GH treatment, the mean height, SH, and foot SDS had normalized (ie, higher than −2.00 SDS), whereas the mean hand, tibia, and armspan had normalized after 1 year of GH treatment. In contrast, anthropometric measurements of the untreated control group remained unchanged or showed only a slight increase during the study period. In the GH group, 3-year changes in height correlated significantly with changes in SH (r = 0.5; P < .001), foot (r = 0.5; P < .001), tibia (r = 0.5; P < .001), armspan (r = 0.3; P < .03), and HC (r = 0.5; P < .001), but not with changes in hand SDS.
SGAL+W Versus SGAL
SGAL+W children had significantly lower values for gestational age, birth length SDS, birth weight SDS, and birth HC SDS compared with SGAL children (Table 3). It is interesting that at baseline, at a mean (SD) age of 5.9 (1.5) years, mean height SDS had become comparable for SGAL+W and SGAL children but the mean HC SDS was still significantly lower in SGAL+W children compared with SGAL children (−1.3 [0.8] vs −0.8 [0.9]; P = .006). At baseline, 14 of 55 SGAL+W children had an HC SDS ≤−2.0 in contrast to 2 of 32 SGAL children (25% vs 6%; P < .001). Also, WH SDS was significantly lower in SGAL+W children (−1.8 [1.0] vs −0.8 [0.9]; P < .001).
Three years of GH treatment induced a significant increase in height, weight, and HC SDS, which was similar for SGAL+W and SGAL children. HC SDS, however, still remained lower in SGAL+W children compared with SGAL children after 3 years of GH treatment, although the difference did not reach level of significance. The 3-year change in HC SDS correlated negatively with age at start of GH treatment in both SGAL+W and SGAL children (r = 0.4; P = .01 and r = 0.7; P < .001). Thus, the younger they started GH treatment, the greater the increase in HC SDS. After 3 years of GH treatment, WH SDS was still significantly lower in the SGAL+W group compared with the SGAL group. In SGAL+W and SGAL control subjects, no changes in height, weight, and HC SDS were found during the 3-year follow-up period. The SDSs for SH, hand, tibia, foot, and armspan did not differ between SGAL+W and SGAL children, neither at start of the study nor after 3 years of GH treatment (data not shown).
Short children who were born SGA show a reduced size of their SH; armspan; and tibia, foot, and hand size, which is in proportion to their reduced height. Children who were treated with GH for 3 years had a normalization of their height in contrast to children who remained untreated for 3 years. In the GH-treated children SH; armspan; and tibia, foot, and hand SDS increased in concordance with their increase in height SDS, indicating that body proportions remained normal during 3 years of GH treatment. Conversely, height and other anthropometric measurements remained unchanged in the untreated control group. Therefore, the present study now demonstrates that the observed changes in various body parts of GH-treated children are the result of GH treatment. Our results agree with previous findings in a comparable group of short SGA children.8 In that study, however, all children received GH treatment and therefore no conclusions could be drawn concerning a causal effect of GH treatment on changes in body proportions. The effects of long-term GH treatment on body proportions are unknown and need additional investigation.
It is interesting that HC of the total group of short SGA children was less affected. Although height and all other anthropometric measurements showed a mean value that was significantly less than −2.0 SDS, the mean HC SDS was the only parameter that showed a value higher than −2.0 SDS. Of a total of 87 short SGA children, only 16 (18%) had an HC SDS below the normal range, ie, less than −2.0 SDS, at a mean (SD) age of 5.9 (1.5) years. It is interesting that most (14 of 16) children with an HC SDS less than −2.00 had been proportionately small at birth (SGAL+W). Although the magnitude of spontaneous catch-up growth after birth in HC had been greater in SGAL+W children compared with SGAL children (1.0 [1.2] vs 0.0 [1.1]; P < .001), HC SDS was still significantly lower in SGAL+W children at a mean age of 5.9 years. This finding strongly suggests that short SGA children who were proportionately small at birth have a greater risk for a relatively small HC during childhood.
During GH treatment, HC SDS increased similarly in both SGAL+W and SGAL children. Because at the start of the study HC SDS was significantly smaller in SGAL+W children, these children still had a smaller HC after 3 years of GH treatment. Apparently, GH treatment has a similar effect on HC in both SGAL+W and SGAL children. In contrast, those who remained untreated for 3 years, being born either as SGAL+W or as SGAL, had no gain in HC SDS during the 3-year study period. Thus, GH treatment induces a significant increase in HC SDS in both SGAL+W and SGAL children.
There has been some debate about the effects of being born SGA on intelligence and behavior.15 Recently, 2 large population-based studies showed lower school performances in 16- and 20-year-olds who were born SGA.10,16 It has also been reported that a lower intelligence was associated with a smaller HC in SGA children at various ages.9,17 Lundgren et al17 found an important association between both birth length and persistent short stature and subnormal intellectual performances. Our first SGA study was the first to investigate psychological functioning in a group of short SGA children who had similar inclusion criteria as the present study.18 This study showed a significant reduction in intelligence and attention and an increase in behavioral and emotional problems. Both intelligence and attention were significantly associated with HC during childhood. Our present study shows that at a mean age of 5.9 years, short SGAL+W children have smaller HC than SGAL children. These findings suggest that SGAL+W children might have a higher risk for lower school achievements and psychological problems. Unfortunately, our present study did not include intelligence tests.
Because GH treatment results in an increase in HC and HC is associated with intelligence, we may speculate that GH treatment could have a positive influence on intellectual performances of short SGA children. Van der Reijden et al18 performed intelligence tests in short SGA children before and after 2 years of GH treatment and found a significant increase in IQ after 2 years. Unfortunately, this study did not include an untreated control group. Our study shows that the gain in HC SDS during GH treatment correlated negatively with age at start. Thus, if GH treatment would stimulate intellectual and psychological development of short SGA children, then it may be important to start treatment at a young age. This would especially apply to SGA children who were proportionately small at birth, because they are particularly at risk for a small HC during childhood.
Gestational age of SGAL+W children was significantly shorter than that of SGAL children. Despite their shorter gestational age, their birth length, birth weight, and birth HC were more severely reduced, suggesting that growth retardation was more severe or occurred early during pregnancy. Possibly, in these SGAL+W children, genetic causes could play a role. Recently, we investigated whether the IGF-I gene was involved in the phenotype of short children who were born SGA.19 In a large group of SGAL+W children, allele 191 of the IGF1.PC1 marker, located between exons 2 and 3 of the IGF-I gene, was transmitted more often from parent to child than one would expect on the basis of Mendelian frequencies. Children who carry this allele had significantly lower serum IGF-I levels and had a smaller HC at the age of 1.3 years compared with children who do not carry this allele. Thus, this functional IGF-I polymorphism might play a role in short SGA children who were born proportionately small. Woods et al20 described severe prenatal and postnatal growth failure as well as mental retardation in a child with a partial deletion of the IGF-I gene. These studies suggest that IGF-I not only plays an important role in somatic growth but also might be involved in the development of neuropsychological functioning. Additional research is required to investigate the role of GH treatment on the increase in HC and its effect on neuropsychological functioning in SGA children.
During GH treatment, no adverse events were detected and glycosylated hemoglobin levels remained normal in all children (data not shown). A relative insulin resistance with an increase in fasting insulin levels has been described during GH treatment.21,22 However, short untreated SGA children already have a reduced insulin sensitivity and an increased risk for the development of type 2 diabetes.23,24 Van Pareren et al25 showed that the GH-induced insulin insensitivity disappeared after discontinuation of GH after long-term GH treatment. However, it remains important to evaluate regularly glucose metabolism during and after discontinuation of GH treatment.
In conclusion, untreated short SGA children generally have normal body proportions. Besides height, the sizes of SH, armspan, tibia, foot, and hand all are significantly lower than −2.0 SDS. In contrast, HC is less severely affected and therefore relatively large in these children (−1.1 SDS). Those who were born SGAL+W still had a smaller HC at the age of 5.9 years compared with those who were born SGAL. Three years of GH treatment induced a proportionate growth resulting in normalization of both height and other anthropometric measurements, including HC.
This study was supported by Novo Nordisk Farma B.V., Netherlands.
E. Lems and I. van Slobbe, research nurses, are greatly acknowledged for assistance.
The participating centers were E.C.A.M. Houdijk, Free University Hospital Amsterdam; J.C. Mulder, Rijnstate Hospital, Arnhem; J.J.J. Waelkens, Catharina Hospital, Eindhoven; R.J.H. Odink, Beatrix Children's Hospital, Groningen; W.H. Stokvis, Medical University Center Leiden; C. Rongen-Westerlaken, Canisius-Wilhelmina Hospital, Nijmegen; A.C.S. Hokken-Koelega, Sophia Children's Hospital, Rotterdam; H.M. Reeser, Juliana Children's Hospital, The Hague; M. Jansen, Wilhelmina Children's Hospital, Utrecht, Netherlands.
- ↵de Waal WJ, Hokken-Koelega AC, Stijnen T, de Muinck Keizer-Schrama SM, Drop SL. Endogenous and stimulated GH secretion, urinary GH excretion, and plasma IGF-I and IGF-II levels in prepubertal children with short stature after intrauterine growth retardation. The Dutch Working Group on Growth Hormone. Clin Endocrinol (Oxf).1994;41 :621– 630
- ↵Tanner JM, Whitehouse RH. Clinical longitudinal standards for height, weight, height velocity, weight velocity, and stages of puberty. Arch Dis Child.1976;51 :170– 179
- ↵Gerver WJM, De Bruin R. Paediatric Morphometrics: A Reference Manual. Utrecht, Netherlands: Bunge; 1996
- ↵Grantham-McGregor SM. Small for gestational age, term babies, in the first six years of life. Eur J Clin Nutr.1998;52(suppl 1) :S59– S64
- ↵Larroque B, Bertrais S, Czernichow P, Leger J. School difficulties in 20-year-olds who were born small for gestational age at term in a regional cohort study. Pediatrics.2001;108 :111– 115
- ↵van der Reijden-Lakeman I. Growing Pains. Thesis. Rotterdam, Netherlands: Erasmus University Rotterdam; 1996
- ↵Rizza RA, Mandarino LJ, Gerich JE. Effects of growth hormone on insulin action in man.Mechanisms of insulin resistance, impaired suppression of glucose production, and impaired stimulation of glucose utilization. Diabetes.1982;31 :663– 669
- Copyright © 2004 by the American Academy of Pediatrics