| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Pediatric Clinical Chemistry |
Departments of1 Pediatrics and2 Public Health Science, La Sapienza University of Rome;3 National Research Council, Rome;4 Villa San Pietro Hospital, Rome;5 National Institute of Health, Rome, Italy.
aAddress correspondence to this author at: Department of Pediatrics, La Sapienza University of Rome, Viale Regina Elena, 324 00161 Rome, Italy. Fax 39-06-4997-9215; e-mail claudio.chiesa{at}uniroma1.it.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: In 153 delivering mothers and their offspring at birth, we prospectively investigated the association between mothers’ and babies’ serum concentrations of ghrelin, leptin, insulin, IGF-1, and IGFBP-3 and neonatal anthropometric characteristics and the growth of the fetus. We also tried to put babies’ serum glucose and GH measurements in this context.
Results: Birth weight (BW), birth length, head circumference, and ponderal index (PI) were positively associated with cord IGF-1, IGFBP-3, and leptin and negatively associated with GH. BW was independently associated with maternal stature and prepartum weight, birth length with maternal stature, PI with maternal insulin and prepartum weight, and head circumference with maternal ghrelin. Compared with preterm infants whose development was appropriate for gestational age (AGA), preterm growth-restricted babies displayed alteration in GH-IGF axis (increased GH and low IGF-1 and IGFBP-3 concentrations), low leptin and glucose concentrations, and increased ghrelin concentrations. In large-for-gestational-age (LGA) babies, leptin, IGFBP-3, insulin, and glucose concentrations were significantly higher in asymmetric LGA newborns than in symmetric LGA and AGA newborns.
Conclusions: We found relationships between metabolic factors, fetal growth, and anthropometry. Intrauterine growth restriction involved alteration in the fetal GH-IGF axis, with relatively low leptin and glucose concentrations and increased ghrelin concentrations. Leptin, insulin, and IGFBP-3 delineated subtypes of fetal overgrowth.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As oral and intravenous glucose have been reported to decrease ghrelin levels (6)(7), only mothers who received normal saline solution during labor were eligible. Mothers with multiple gestation and clinically evident intraamniotic infection were excluded, as were mothers who received hormone replacement therapy during pregnancy and those whose fetuses presented chromosomal abnormalities or abnormal anatomy. Thus 165 mothers were consecutively enrolled.
Newborns were categorized at birth as appropriate for GA (AGA), large for GA (LGA), and small for GA (SGA) according to Italian standards for BW, sex, and GA (8). Since there is no consensus as to the BW threshold to define a macrosomic neonate (9), defining macrosomia as LGA circumvents this problem. The type of macrosomia was established on the basis of ponderal index (PI) and Miller charts (10): LGA newborns whose PI was above the 97th percentile were classified as having asymmetric macrosomia (10)(11), and LGA newborns whose PI was between the 10th and 90th percentiles were classified as symmetric (10)(11).
The mother-infant pair was excluded from the analysis if (a) the newborn had congenital or perinatal infection; (b) the newborn required intensive care treatment in the immediate postnatal period (n = 6); (c) the baby’s small size for GA at birth conflicted with fetal biometric results from antenatal surveillance (n = 1); or (d) mothers’ and babies’ serum samples were not tested for all study parameters (n = 5). Thus 153 pairs qualified for analysis.
assays
We assessed immunoreactive ghrelin levels by RIA as reported (12). We also used RIA to measure human leptin (DRG Diagnostica) and insulin (CIS Bio International, Schering S.A.). We used an IRMA to measure IGF-1 (Immunotech, Beckman Coulter, Inc.), GH (CIS Bio International, Schering S.A.), and IGFBP-3 (13) (DRG Diagnostica). We measured glucose concentrations by the glucose oxidase method using a Synchron LX 20 (Beckman Coulter Inc.).
statistical analyses
For all the variables studied, we constructed frequency distributions and calculated summary descriptive measures, such as means and proportions. The distributions of the hormone levels, the duration of labor, and the time from the rupture of the membranes to delivery were not symmetric but had a long tail to the right. These positively skewed variables were transformed using natural logarithm before their use in subsequent analyses. We used 1-way ANOVA to compare variables in the term AGA infants and asymmetric and symmetric LGA infants and 2-sample t tests to compare variables between preterm AGA and growth-restricted infants.
We used multiple regression analysis to investigate the effects of the hormones on neonatal anthropometric indices. In a simple multiple regression of a quantitative variable y on a quantitative variable x, the interpretation of the observed regression coefficient b is the estimated average increase in y when the value of x increases by 1 unit. However, suppose the equation obtained from the regression analysis were: y = a + b (ln x). If y1 denotes the value of y that corresponds to a particular value of x, say x1, y1 = a + b (ln x1). If x1 increases by p% = 10%, the resulting value of y would be y2 where y2 = a + b (ln 1.1x1) = a + b (ln 1.1) + b ln x1 = a + 0.09531b + b ln x1. That is, 0.09531b is the average increase in y associated with a 10% increase in x. For example, from the results of the regression of birth weight in the first row of Table 2
, the coefficient for the independent variable, ln ghrelin, is b = 53. This implies that on average, when ln maternal ghrelin increases by 1 (that is ghrelin is multiplied by 2.72), birth weight increases, on average, 53 g. If ghrelin increases by 10%, on average, birth weight increases by 0.09531 x 53 = 5.1 g. In the case in which the independent variable is log time, instead of calculating the percentage increase associated with a 10% increase in duration of labor or time since rupture of membranes, the results show the percentage increase in hormone level associated with a doubling of the time.
|
For the comparisons between preterm babies with and without IUGR, the sample sizes were 28 and 11, respectively. These were not calculated a priori for the hypotheses to be tested, but were the number of eligible mother-infant pairs that were observed in the study period. To compare the level of ghrelin and leptin in cord blood, the power can be calculated a posteriori, using the results obtained from the study to estimate the standard deviations. The calculated power to detect a difference of 20% in geometric mean hormone levels (significantly at the 5% level) between babies with and without IUGR was 10% for ghrelin and 12% for leptin. The corresponding powers to detect the differences actually observed are 98% and 87%, respectively. Subgroup analyses are based on small samples, and consequently lack of statistical significance does not imply the absence of an association.
To achieve the objectives of this study, a very large number of statistical significance tests are required. This raises the serious problem of multiple testing. There are techniques that can reduce the probability of type I errors, for example, the Bonferroni correction. However, this study is explorative in the sense that many comparisons are made simply to discover if there seems to be an association or not, and for this reason, no corrections have been applied. This means that the calculated levels of significance for individual comparisons should be interpreted with caution. Indeed, the significant results should be considered as "hypothesis generating" rather than "hypothesis testing."
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
presents anthropometric and clinical characteristics of the 153 mother-newborn pairs, who were almost all (97.4%) Italian.
|
Paternal weight and height were not significantly associated with any anthropometric measure of the newborn. In contrast, BW and PI increased, independently of GA and sex, on average 9.7 g (95% CI 2.1–17.2, P <0.01) and 0.5% (0.1–0.9, P <0.05), respectively, per kg of maternal prepartum weight. Also, BW and body length increased, independently of GA and sex, on average 17 g (4.3–27, P <0.01) and 0.08 cm (0.03–0.14, P <0.01), respectively, per cm of maternal height.
Maternal leptin, IGFBP-3, and IGF-1 were not associated with any anthropometric measure of the newborn. In contrast, maternal insulin was positively associated with PI and tended to correlate with BW (P = 0.06), whereas ghrelin was positively associated with head circumference (Table 2
).
Babies’ ghrelin and insulin were not associated with any anthropometric measure at birth. In contrast, babies’ IGF-1 and IGFBP-3 were positively associated with all anthropometric indices; babies’ leptin was positively associated with BW and PI and tended to correlate positively with body length (P = 0.05); and GH was negatively associated with BW and body length (Table 3
).
|
iugr and metabolic and hormonal correlates
IUGR was associated in the fetal compartment with alteration in GH-IGF axis (relatively increased GH and low IGF-1 and IGFBP-3 concentrations), relatively low leptin and glucose concentrations, and increased ghrelin values (Table 4
). Maternal IGF-1, but not ghrelin, leptin, IGFBP-3, or insulin, was associated with IUGR; maternal mean IGF-1 concentrations were significantly lower in pregnancies with IUGR (134 [99–181] µg/L, P <0.05) compared with (preterm) AGA pregnancies (200 [181–221] µg/L).
|
lga newborns: subtypes and clinical, metabolic, and hormonal correlates
Symmetric LGA babies were significantly taller than asymmetric LGA and control AGA neonates (Table 5
). Mothers of 1 of the 10 asymmetric LGA newborns and none of the 17 symmetric LGA newborns had a history of gestational diabetes.
|
GH, IGF-1, and ghrelin concentrations were similar in the 3 groups. However, mean leptin, IGFBP-3, insulin, and glucose concentrations were significantly higher in asymmetric LGA newborns than in symmetric LGA and AGA newborns (Table 5
). Also, women who gave birth to either symmetric or asymmetric LGA newborns did not differ in mean ghrelin, leptin, IGF-1, IGFBP-3, or insulin concentrations vs women delivering term AGA newborns.
maternal and perinatal confounders
Maternal mean ghrelin and insulin increased by 7.4% (95% CI 1.0%–13.8%, P <0.01) and 6.5% (1.0%–11.6%, P <0.01), respectively, per week of GA at delivery, whereas maternal IGF-1 decreased by 9.3% (8.9%–9.7%, P <0.001) per week of GA. Women with history of prenatal steroids had lower mean insulin levels than those without (12.4 [8.7–17.6] vs 18.4 [16.3–20.7] mU/L; P <0.01), but higher IGF-1 (236 [178–313] vs 178 [161–195] µg/L; P <0.05). Women with gestational hypertension had higher leptin concentrations than those without (19.6 [13.2–28.9] vs 12.4 [10.9–14.0] µg/L; P <0.01). Women with intrapartum fetal distress (IFD) had higher leptin than those without (17.1 [13.7–21.4] vs 12.1 [10.5–13.9] µg/L; P <0.01), as well as of IGF-1 (252 [205–311] vs 169 [153–187] µg/L; P <0.0001) and IGFBP-3 (6576 [5824–7425] vs 5504 [5037–6014] µg/L; P <0.05), but lower ghrelin (90 [70–115] vs 138 [118–161] ng/L; P <0.01). Women who delivered by elective cesarean section had lower leptin (11.0 [9.9–11.2] µg/L), IGF-1 (164 [148–181] µg/L), and IGFBP-3 (4914 [4447–5431] µg/L) than those who delivered by spontaneous vaginal delivery (16.4 [13.5–20.0] µg/L, P <0.01; 221 [164–299] µg/L, P <0.05; and 7331 [6003–8955] µg/L, P <0.01, respectively) or urgent cesarean section (20.0 [14.9–27.1] µg/L, P <0.0001; 221 [181–270] µg/L, P <0.01; and 6003 [5431–7634] µg/L, P <0.05, respectively). Maternal mean leptin increased by 6.8% (2.7%–11.1%, P <0.01) for each doubling of duration of active labor and by 6.4% (2.0%–10.4%, P <0.01) for each doubling of time of ruptured membranes. Likewise, maternal mean IGF-1 increased by 2.9% (0.7%–5.1%, P <0.01) for each doubling of duration of active labor and by 4.6 (1.4%–7.6%, P <0.01) for each doubling of time of ruptured membranes.
Compared with male babies, female babies had higher mean leptin (8.0 [6.6–9.6] vs 5.2 [4.4–6.2] µg/L; P <0.001) and IGFBP-3 (2113 [1956–2282] vs 1790 [1652–1940] µg/L; P <0.01). At birth, mean ghrelin, leptin, IGF-1, and IGFBP-3 increased by 13.0% (6.0%–18.0%, P <0.0001), 15.0% (10.0%–21.0%, P <0.0001), 12.2% (4.3%–20.2%, P <0.01), and 5.0% (2.5%–7.4%, P <0.0001), respectively, per week of GA at delivery. Babies with prenatal exposure to steroids had lower levels of ghrelin than those without (242 [180–326] vs 462 [395–540] ng/L; P <0.001), as well as of leptin (3.5 [2.3–5.2] vs 7.2 [6.4–8.2] µg/L; P <0.0001), IGF-1 (57 [44–76] vs 92 [84–100] µg/L; P <0.0001]), and IGFBP-3 (1404 [1195–1651] vs 2071 [1961–2187] µg/L; P <0.0001). Mean ghrelin concentrations were higher in babies of smoking mothers than in those born to nonsmokers (544 [403–735] vs 365 [330–445] ng/L; P <0.01). Babies of mothers with gestational hypertension tended to have lower IGF-1 concentrations than those born to mothers without (69.7 [51.5–94.4] vs 88.3 [80.7–96.8] µg/L; P = 0.06). Babies with IFD had lower ghrelin (290 [213–394] vs 464 [398–545] ng/L; P <0.01) and IGF-1 (72.3 [57.5–91.0] vs 90.0 [82.0–99.0] µg/L; P <0.05). Babies born by elective cesarean section had higher ghrelin (492 [446–544]) than those born by spontaneous vaginal delivery (298 [181–492]; P <0.01) or urgent cesarean section (330 [221–492] ng/L; P <0.05). Babies’ mean IGF-1 decreased by 2.2% (0.7%–4.2%, P <0.05) for each doubling of time of ruptured membranes.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Epidemiologic studies suggest that maternal diet and metabolic function can influence the pattern of human fetal growth (15). Similarly, we found that maternal prepartum weight correlated positively with baby’s PI and BW, and maternal insulin was positively associated with baby’s PI and tended to positively correlate with BW.
Reduced maternal IGF-1 has been described in cases of IUGR (16); however, other studies failed to demonstrate such an association (16). We found no association between maternal IGF-1 and BW within the entire study population. When the analysis was confined to preterm newborns, however, we found that maternal IGF-1 was significantly lower in those who gave birth to neonates with IUGR. Possible explanations for this include a secondary consequence of placental dysfunction or an adaptation to limit glucose supply to a hypoxic fetus (17).
Our study also found that maternal ghrelin concentration was positively associated with baby’s head circumference (and therefore brain development). This is consistent with recent experimental findings suggesting that maternal ghrelin is easily transferred to the fetal circulation, and then prompts fetal growth through direct stimulation of cell proliferation in the second half of pregnancy (18). Recent reports that ghrelin directly stimulates bone formation (19) also support this hypothesis. Thus, a better understanding of the role of maternal ghrelin in the development of various fetal tissues may also help explain the mechanisms that regulate fetal growth.
iugr
In this and other studies (20), diminished intrauterine growth involved alterations in the IGF axis. If a fetus grows in an environment of relative deprivation, its hormonal milieu adapts to the relative scarcity of nutrients. Poor growth may then be regarded as one such adaptation, and the alteration in GH-IGF axis, with relatively low leptin and glucose concentrations and increased ghrelin concentration, reflects these conditions (20). Our observation that ghrelin was also increased in infants born to mothers who smoked in pregnancy provides further evidence that ghrelin may play a role in fetal adaptation to intrauterine malnutrition (21). Tobacco use is more closely linked to restricted fetal growth than to preterm birth (22). It may be of interest that ghrelin values are higher in patients with anorexia nervosa than in controls, and that anorexia nervosa is also associated with high concentrations of GH and low concentrations of IGF-1, suggestive of a nutritionally acquired lack of GH action or GH resistance (1).
the large newborn
Symmetric macrosomia is a result of normal uterine environment that stimulates growth and includes normal, genetically large infants. The fetus may be large in size but not distinguished by any abnormalities (14). In our study, symmetric LGA newborns and control AGA newborns had similar PI and metabolic parameters. However, they were taller than control and asymmetric LGA newborns, providing indirect evidence of a genetic cause in symmetric fetal overgrowth. In contrast, asymmetric macrosomia is associated with accelerated fetal growth (14). In addition to weight, neonatal body proportions have a role in defining this type of macrosomia. Fetuses with asymmetric macrosomia have a greater shoulder circumference to head circumference ratio than similar-weight fetuses with symmetric macrosomia. The extra fat in these infants, therefore, may be concentrated in the upper body. Disproportionate macrosomia is also characterized by organomegaly and should be considered a pathologic entity. In our study, compared with symmetric LGA and AGA newborns, asymmetric LGA newborns had higher insulin, leptin, and IGFBP-3 concentrations and lower glucose concentrations. These data suggest that asymmetric fetal macrosomia might be caused by an abnormal intrauterine environment.
Although the risk of asymmetric macrosomia is high in pregnancies of diabetic women, our asymmetric LGA neonates with hyperinsulinemia and hypoglycemia were almost entirely born to nondiabetic mothers. This is consistent with previous reports (11)(23)(24). The Pedersen hypothesis proposes that the cause of macrosomia in diabetic pregnancies is fetal hyperinsulinism secondary to an excessive supply of substrate due to maternal hyperglycemia (25). Our data support the view that hyperinsulinemia in fetuses whose mothers do not have gestational diabetes or confirmed impaired glucose tolerance might be attributed to mild maternal hyperglycemia below the diagnostic threshold (26). Even a limited degree of maternal hyperglycemia, still considered to be in the normal range, may affect fetal weight (23).
In vivo and in vitro studies have demonstrated a stimulatory effect of insulin on leptin synthesis (27). Because IGFBP-3 is also regulated by insulin (28), our results show an interplay between leptin, insulin, and IGFBP-3 (the principal carrier of the circulating IGFs) as positive modulators of fetal growth.
In our study, no significant differences in IGF-1 concentrations between the two types of macrosomia or between LGA and AGA newborns were observed. It is possible that in some hyperinsulinemic conditions, insulin itself exerts a direct effect on fetal growth by binding to the IGF-1 receptor (29). The structural homology between the insulin receptor and the IGF-1 receptor makes this interaction feasible. On the other hand, insulin can also regulate the IGF axis by control of IGFBP expression. An inverse control of IGFBP expression by insulin may regulate IGF bioavailability to high-affinity receptors, and thereby fetal growth (30).
maternal and perinatal confounders
A novel finding is that both maternal and fetal ghrelin increase with length of gestation at delivery. This is relevant given the recent experimental findings that suggest an important role of maternal ghrelin in rat fetal development. Nakahara et al. (18) demonstrated that the placenta contributes to the circulating pool of maternal ghrelin during late gestation, and that maternal ghrelin rapidly and easily crosses to the fetus. Also, they showed that while fetal ghrelin originates from the maternal placenta and/or maternal blood, acyl and des-acyl ghrelin are still present in the maternal and fetal circulations during the second half of pregnancy. They also demonstrated that exogenous chronic treatment of the mother with ghrelin increases fetal BW, whereas mothers immunized against ghrelin deliver fetuses with a lower BW (18). Taken together, our data and these findings indicate a role of maternal and fetal ghrelin in fetal development.
Another novel finding is that while maternal IGF-1 correlated negatively with length of gestation, fetal IGF-1 correlated positively, which may imply a feedback mechanism or physiological process of maternal constraint of normal fetal growth. Maternal constraint refers to the limited capacity of the uterus to support fetal growth and is important to limit fetal overgrowth. Maternal constraint may limit supply by maternal size or nutrient availability (16).
As expected (31), delivering mothers with a history of gestational hypertension had higher leptin levels. Hypertensive disorders of pregnancy cause arteriolar vasoconstriction, which leads to placental hypoxia. The probable role of placental hypoxia in the increased production of leptin has been reported by Mise et al. (32). Yet, as it has been shown that acute fetal hypoxia reduces IGF-1 levels in the ovine fetus (33), our novel finding of reduced cord IGF-1 levels in cases with gestational hypertension supports that hypothesis. Consistent with the above observations (33) is also our novel finding of reduced cord IGF-1 in the presence of IFD.
Not surprisingly (34)(35), IFD, modes of delivery that elicit stressful stimuli, and prolonged active labor were found to increase maternal leptin and decrease maternal ghrelin. Of note, however, is the novel finding that babies’ ghrelin decreased in the presence of IFD. This might represent an attempt to antagonize the signals evoked during inflammatory stress, perhaps by proinflammatory factors such as leptin (36) or other cytokines.
In agreement with recent observations (36)(37), delivering mothers with a longer time of ruptured membranes had increased serum leptin and IGF-1 concentrations. An interesting finding, however, was that babies born to mothers with prolonged ruptured membranes had lower IGF-1. Relevant to this finding, inflammatory reactions have been reported to reduce serum IGF-1 (and IGFBPs) (38).
In summary, in this study we found relationships between metabolic factors, fetal growth, and anthropometry. IUGR involved alteration in the fetal GH-IGF axis, with relatively low leptin and glucose concentrations and increased ghrelin concentrations. Leptin, insulin, and IGFBP-3 delineated subtypes of fetal overgrowth. Maternal and perinatal factors should be taken into account to optimize our understanding of the mechanisms by which endocrine factors may regulate fetal growth.
|
Acknowledgments |
|---|
Financial Disclosures: None declared.
Acknowledgments: We are indebted to Professors Alessandra Panero and Marco Matrunola, without whose teachings and encouragement this study could not have come to fruition.
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  F. Darendeliler, S. Poyrazoglu, F. Bas, O. Sancakli, and G. Gokcay Ghrelin levels are decreased in non-obese prepubertal children born large for gestational age Eur. J. Endocrinol., June 1, 2009; 160(6): 951 - 956. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
