Main

Nothing is known about the predictive value of neonatal parameters on the catch-up growth that occurs in most but not all IUGR children during the first 2 y of life. Previous studies have shown an increase in GH levels(1, 2) and low(3–6), normal(7), or high(8–10) serum IGF-I levels in IUGR neonates compared with those found in appropriate weight for age neonates. Only one longitudinal study of infants up to 1 y of age demonstrated lower IGF-I levels in a group of five IUGR infants whose length was less than -2 SD at 12 mo of age. Thus, the low IGF-I level might reflect poor postnatal growth(8). Although a decrease in the amount of IGFBP3 was found in cord sera from human neonates with IUGR, the serum concentration of GH-dependent IGFBP3 at birth and during the catch-up growth period has never been described in cases of IUGR.

To elucidate any hormonal abnormalities at birth due to IUGR and to identify any parameters which could be useful in predicting later growth, we measured the GH, IGF-I, and IGFBP3 levels in the cord blood of both preterm and full-term newborn infants showing IUGR, and we followed their serum GH, IGF-I, and IGFBP3 levels as well as their auxologic status during the first 24 mo of life.

METHODS

Study population and protocol. Neonates born with IUGR in eight maternity hospitals during an 18-mo period (years 1990-1991), were recruited in the study. All IUGR subjects were defined as having a birth weight below the 3rd percentile according to the French Leroy's growth curve standards(11) were included in the study. Malformation and/or chromosomal abnormalities or maternal short stature (≤ -2 SD) were not considered as exclusion criteria. A total of 317 IUGR subjects were recruited and studied at birth, at 3 d, and at 1, 3, 6, 12, 18, and 24 mo of life. Blood samples were taken from the cord at birth and at each visit thereafter. The control population (obtained transversally) for the biologic parameters were normal, healthy, age-matched infants with a birth weight appropriate for gestational age (>25th percentile), n = 84 at birth from the cord blood (48% preterm newborn) with auxologic data (as indicated inTable 1), n = 36 at 3 d, n = 51 at 1 mo, n = 16 at 3 mo, n = 12 at 6 mo (term newborn only), and n = 37 at 12 and 24 mo of life.

Table 1 Characteristics of the neonates at birth
Full size table

They were subjected to blood sampling for routine examinations at 3 d. Euthyroid children, false positive cases in our screening program, served as the control group for the 1-, 3-, and 6-mo-old children, and normal children, who undergo blood sampling for free health examination, served as the control group for the 12- and 24-mo-old children.

Procedures. From birth until 6 mo of age, height, weight, and head circumference were expressed as a SDS and corrected for gestational age and sex according to the intra- and extra-uterine growth standard of Largoet al.(12). From 6 to 24 mo the Sempe growth standard(13) was used.

Supine length (height) measurements required two observers and were made twice at each time of recording with the infants, with the head on the midline and the knees extended, on a neonatometer (Harpenden produced by Holtain Ltd.).

The asymmetric IUGR group was defined as having a birth weight of <-2SDS but height > -2 SDS and the symmetric IUGR group as being < -2 SDS for both body weight and height according to the gestational age at birth.

To evaluate the intrauterine nutritional state, the PI was calculated as the ratio of birth weight (g) to the cube of the length (cm3) × 100 corrected for gestational age according to Miller's standards(14). Two groups were constituted according to a PI ≤ or >3rd percentile at birth.

At 1 and 2 y of age, weight for height was expressed as body mass index(weight (kg)/height (m2)) in SDS for chronologic age(15). Historic high risk factors and high risk factors that developed during gestation were recorded at birth during the recruitment.

Hormone assays. Blood samples were collected during the study at any time of the day, and the sera were stored at -20 °C until assayed. The serum GH concentration (ng/mL) was measured on a single sample by RIA using a solid phase two-site immunoradiometric assay (Elsa-hGH, CIS bio International). Standards were calibrated against as international standard(1st IRP 66/217).

IGF-I (ng/mL) was measured by RIA after separation from IGFBP by acid chromatography following a published procedure(16). The polyclonal IGF-I antiserum used in the RIA was kindly provided by P. Chatelain(Lyon, France). The within-run coefficients of variation were 5.5 and 8.8%, respectively for the middle (137 ng/mL) and low values (85 ng/mL), and the between-run coefficients of variation were, respectively, 12.4, 9.4, and 15.3% for the high (from 250 ng/mL), middle (from 125 ng/mL), and low values (from 62.5 ng/mL) (concentration values after dilution).

IGFBP3 (μg/mL) was measured by RIA using a commercially available kit(Mediagost)(17). The mean within and between run coefficients of variation were 3.5 and 6.5%.

Statistical analysis. All results were expressed as the mean± SD. Statistical analysis was performed using paired and unpairedt tests and the Mann-Whitney U test. Correlations between variables were assessed using linear regression analyses. Comparisons of group mean parameter estimates were made using analysis of variance. Multivariate linear regression analysis was used to study differences in hormonal levels and child growth status (IUGR or appropriate for gestational age) adjusted for gestational age. SAS (Statistical Analysis System, Cary, NJ) software running under OS2 on a compaq Deskpro 486/33 M microcomputer was used.

The study was reviewed and approved by the faculty ethical committee. It was explained to each parent who signed a written consent.

RESULTS

Description of the cohort. Characteristics of the IUGR and control newborn infants are indicated in Table 1. Among the 317 IUGR newborns included in the study, 254 IUGR subjects (female, 149; male, 105) were studied for hormonal evaluation at birth. The mean gestational age was 37.6 ± 2.5 wk, ranging from 29 to 42 wk, and 32% were preterm newborn (<37 wk gestational age). Factors associated with IUGR included pregnancy-induced hypertension (n = 67), multiple pregnancies(n = 38), smoking (n = 95), heavy alcohol (n = 7) consumption, congenital anomalies or chromosomal abnormalities (n= 26), preeclampsia (n = 8), and maternal short stature (<152 cm)(n = 13).

Medical indications of risk associated with growth status at birth such as low Apgar score (≤ 7 at 1 min of life) (n = 60) were assessed. Among the 254 IUGR infants, 53% were symmetrically IUGR for height and weight. The PI was normal (>3rd percentile) for 62% and low (≤3rd percentile) for 38% of the infants.

An auxologic description of the entire cohort and part of that cohort only in those cases submitted to hormonal evaluations is reported inTable 2, and the number of cases of each group are indicated. Not all biologic parameters and growth measurements were available on every child at each time because of noncompliance of the families either in the blood sample collection and/or in the long-term follow up. No significant differences among these groups were found for gestational age, the known causes of IUGR, and the auxologic data either at birth or during follow-up. Short stature (height ≤ -2 SDS) was observed in 8% of the subjects of the entire cohort studied at 2 y of age.

Table 2 Mean (±SD) growth parameters expressed as SDS from 1 to 24 mo of life in IUGR children, studied for the entire cohort(A) and for patients evaluated for hormonal parameters (B)
Full size table

Serum GH, IGF-I, and IGFBP3 levels at birth. The individual serum GH, IGF-I, and IGFBP3 levels at birth according to the gestational age are indicated in Table 3 and in Figure 1, A, B, and C. The control group showed a significant decrease of serum GH levels (r = -0.43, p = 0.0001) and a progressive increase of serum IGF-I levels (r = 0.46, p = 0.0001) during the third trimester of pregnancy when results at birth were considered as a function of gestational age. No significant change in serum IGFBP3 levels was observed as a function of gestational age. In the IUGR group, no changes were found in serum GH, IGF-I, or IGFBP3 levels during the third trimester of pregnancy, when results at birth were considered as a function of gestational age.

Table 3 Mean (±SD) serum GH, IGF-I, and IGF BP3 levels in IUGR and control children from birth to 24 mo of age
Full size table
Figure 1
figure 1

(A, B, and C) Serum GH, IGF-I and IGFBP3 levels at birth according to the gestational age and growth status in IUGR (•) compared with normal neonates (○) [linear regression lines were shown with 95% CL for mean (- - - - - -); — for intrauterine growth-retarded neonates, - for normal neonates]. Multivariate linear regression analysis was made on log values. Log GH = 4.37 - 0.037 gestational age + 0.63 growth status. F = 26.95, r = 0.37, p= 0.0001. Log IGF-I = 1.86 + 0.06 gestational age - 0.996 growth status.F = 63.86, r = 0.53, p = 0.0001. Multivariate model on log binding protein 3 values was not significant as no relation with gestational age was found. Serum IGFBP3 were significantly reduced in case of IUGR (F = 14.51, r = 0.32, p = 0.0001).

Full size image

Multivariate linear regression analysis, made on log values, demonstrated significantly higher serum GH levels (r = 0.37, p = 0.0001) and lower serum IGF-I levels (r = 0.53, p = 0.0001) in IUGR neonates compared with controls according to gestational age. Serum IGFBP3 levels were significantly lower in IUGR as compared with controls(r = 0.32, p = 0.0001) with no relation to gestational age.

Wide individual variations were observed for all hormonal parameters in IUGR and control groups. In IUGR infants a significant correlation was found between serum IGF-I levels and birth weight (r = 0.20, p = 0.001). Separate analysis of the hormonal data from neonates for each of the known causes of IUGR, for different sexes, and for the asymmetricversus the symmetric for height group, revealed no differences for the hormonal parameters at birth.

When considered as a function of gestational age, serum IGF-I and IGFBP3 levels were significantly lower in cases where the PI was below the 3rd percentile than in those in which the PI was ≥3rd percentile (IGF-I,p = 0.004; IGFBP3, p = 0.01). For all the subjects(premature and term infants), the mean serum IGF-I level was, respectively, 26± 18 ng/mL versus 35 ± 23 ng/mL and the mean serum IGFBP3 level was 1.0 ± 0.6 μg/mL versus 1.3 ± 1.0μg/mL for neonates with PI < 3rd percentile compared with neonates with PI ≥ 3rd percentile. Serum IGF-I levels were also significantly lower in cases showing a low Apgar score at 1 min of life (≤7) than in those showing an Apgar score >7 (p = 0.0001) with, respectively, a mean serum IGF-I level at 23 ± 17 versus 34 ± 22 ng/mL.

Serum GH, IGF-I, and IGFBP3 levels after birth and up to 2 y of age. The mean serum GH, IGF-I, and IGFBP3 levels in IUGR and control children are presented in Table 3. At 3 d of life, no significant difference was found in the mean serum GH levels of the IUGR group compared with the control group (39 ± 31 ng/mL versus 29± 17 ng/mL, premature infants excluded). The serum IGFBP3 levels were not available at 3 d of life, but attained normal levels at 1 mo of age in the IUGR group. At 3 d of life, the mean serum IGF-I level was significantly lower in the IUGR group than in control children (29 ± 20 ng/mLversus 66 ± 36 ng/mL (p = 0.0001), premature infants excluded). The mean serum IGF-I levels increased progressively during the 1st mo of life, and were normal in the IUGR group at 1 mo of age (79± 33 ng/mL versus 90 ± 35 ng/mL (p = 0.06), premature infants excluded). As shown in Table 3, all biologic parameters had attained normal levels up to 24 mo of life.

As shown in Table 4, when data are analyzed in IUGR children whose height was ≤ -2 SDS compared with IUGR children whose height was > -2 SDS at 2 y of age, the mean serum IGF-I level was found to be significantly reduced only in children with short stature at 2 y of life (50± 18 versus 101 ± 43 ng/mL; p = 0.03).

Table 4 Mean ± SD serum GH, IGF-I, and IGFBP3 levels according to auxologic data in children whose height was ≤-2 SDS as compared with children whose height was >-2 SDS at 2 y of age
Full size table

At 1 and 2 y of age, no significant difference in hormonal parameters was observed as a function of low (≤ -2 SDS) or normal (> -2 SDS) body mass index, which represents an index of nutritional state in postnatal life. A correlation was found between serum IGF-I and IGFBP3 levels at each studied period (at birth: r = 0.52, p = 0.0001; at mo 24:r = 0.59, p = 0.0001) and between serum GH and IGFBP3 levels at birth (r = 0.41, p = 0.0001), and at mo 1(r = -0.26, p = 0.0006 negative correlation). No statistically significant correlation was found between GH and IGF-I levels at any age.

At 1 and 3 mo of age, weight gain was found to be correlated with an increase in serum IGF-I (respectively, r = 0.30, p = 0.0001, at mo 1 and r = 0.28, p = 0.002, at mo 3 for the IGF-I increase in relation to weight gain (Fig 2)). At mo 1, the increase in height was found to be correlated with the serum IGFBP3 increase (r = 0.30, p = 0.004). At any other time studied, the biologic parameters were related neither to length or weight gain nor to the presence or absence of catch-up growth for weight and height. None of the biologic parameters (nor their change) was predictive for later growth.

Figure 2
figure 2

Correlation between the increase in the serum IGF-I levels and weight gain during the first 3 mo of life in IUGR children (linear regression line with 95% confidence level for the mean; gain IGF-I = 18.26 + 12.78 weight gain SDS. F = 10.09, r = 0.28, p = 0.002).

Full size image

DISCUSSION

This study demonstrates low serum IGF-I and IGFBP3 levels and high serum GH levels at birth in a large group of IUGR neonates compared with normal neonates for all gestational ages studied. Our results in normal infants are in agreement with the reported changes in IGF-I, IGFBP3, and GH concentrations with gestational age in serum obtained in utero from normal human fetuses during the second half of gestation(18, 19).

However, wide individual variability was observed in the hormone levels in both normal and IUGR fetuses and neonates. The neonatal hypersomatotropism appears to be characterized by high amplitude pulsatility of serum GH levels(20) which could contribute to the large variation in the early GH results observed. The wide variability in serum IGF-I levels might be partially explained by the hypothesis that IGF-I levels may be genetically influenced as was shown recently in twin children. It was demonstrated that the within-pair correlation of circulating IGF levels for monozygotic twin pairs was significantly higher than that for same sex dizygotic pairs(21).

Previous studies in normal human neonates have yielded a correlation between cord serum IGF-I levels, gestational age, and birth weight(22, 23). Reduced IGF-I levels have also been shown in cord serum from human IUGR neonates and more recently in human fetuses(3–6, 18). Our study is in accordance with these results and documents a decrease of cord serum IGFBP3 levels measured by RIA in IUGR compared with normal neonates.

In animals, fetal growth retardation resulting from maternal dietary protein deprivation(24, 25), uterine artery ligation(26, 27), fetal pancreatectomy(28), or surgical reduction in placental mass(29) induced a decrease in fetal serum IGF-I concentrations. Futhermore, a decrease in fetal IGF-I gene expression and an increase in fetal IGFBP1 gene expression have been demonstrated(25, 27). In cord sera from human fetuses with IUGR, an increase in the amount of IGFBP1 and IGFBP2 have been observed in parallel with a marked decrease in the amount of IGFBP3(6, 30). This demonstrates that the IGF-binding protein which modulates IGF activity, may play a major role in the regulation of fetal growth in relation with the growth and the metabolic status of the fetus, which in part reflects the efficacy of transplacental nutrient transfer and placental perfusion.

Certain authors have reported high(1, 9) or normal(7) IGF-I levels in IUGR infants. These findings are in marked contrast with those of the majority of other investigators and concern very small groups of infants observed at birth or during the first days of life. The reason for this discrepancy is not obvious.

High serum GH levels in IUGR infants were found in this study in agreement with other authors(1, 2, 10). These high circulating GH levels could be secondary to a lessened feedback stimulation by IGF-I related to a diminished sensitivity of the somatotroph cells to the inhibitory actions of IGF-I or a pituitary GH resistance or immaturity of the GH receptor. If low serum IGF-I levels are related to GH resistance, this effect wanes rather quickly, because at the end of the 1st mo of life, serum IGF-I and IGFBP3 levels were normalized. To go further in the discussion of GH resistance in these subjects, measurement of the circulating form of the GH receptor (GH-binding protein) would be useful. This aspect was not studied in the present report. However, low serum GH binding protein levels have been reported in human infants(31). More precise measurement of serum GH-binding protein levels is needed to study their respective values in IUGR compared with appropriate weight for age neonates.

There was no relationship between hormone levels and any given etiology of IUGR, although specific changes in IGF-I and GH levels have been reported in newborn children of alcohol-abusing or smoking mothers(32, 33). Lowered serum IGF-I and IGFBP3 levels were found in IUGR infants showing a low PI, suggesting that intrauterine nutrition influences the levels of these parameters. We can speculate that deficient substrate transfer to the fetus for any given reason might be responsible for both depressed IGF-I synthesis and subsequent growth retardation.

During early neonatal life, serum IGF-I levels appear to be regulated by nutritional factors, because an increase in serum IGF-I levels during the first 3 mo of life was found to be correlated with weight gain and therefore with the increased nutrient intake after intrauterine nutrient deprivation. These findings are in accordance with data from children and adults which demonstrate that circulating IGF-I is markedly influenced by substrate supply, falling during starvation and rising rapidly upon refeeding(34). The results of such studies strongly suggest that nutritional intake is of great importance in the regulation of IGF-I synthesis.

No information was previously available regarding the effects of changes in auxologic status on serum IGF-I, IGFBP3, and GH levels in the perinatal period. Nevertheless no predictive value of neonatal biologic parameters was demonstrated on the extent of later growth. All of these hormonal parameters had attained normal values after the 1st mo, and they remained thus up to 2 y of age. A small number of children with height ≤ -2 SDS at 2 y of age were analyzed for hormonal parameters, and their serum IGF-I levels were lower than those of children whose height was > -2 SDS at 2 y of age. This finding is consistent with a report showing reduced serum IGF-I levels in a group of IUGR children with short stature at the age of 7.5 y compared with controls(35).

Our study is the first to analyze the evolution of serum IGFBP3 levels during the 1st mo of life in normal children. We demonstrated low IGFBP3 levels in cord serum in appropriate weight for age neonates compared with levels later in infancy as has been reported in neonatal rat serum(36). There was a significant correlation between serum IGF-I and IGFBP3 levels throughout the studied period, in agreement with results found later in life in a large group of normal, short, and GH-deficient children(37, 38).

This study allows a better description of the endocrine status in neonates and during the first 2 y of life in IUGR and appropriate weight for age children. Abnormal levels of serum IGF-I and IGFBP3 were related to nutritional factors control ling growth.