|
|
||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Perinatal Center, Institute of Neuroscience and Physiology (NJ, AN, TLP, and TJ), the Department of Obstetrics and Gynecology (MW), and the Department of Clinical Nutrition (LR-H), Sahlgrenska Academy at University of Gothenburg, Göteborg, Sweden; the Department of Informatics, University West, Trollhättan, Sweden (MG); and the Department of Obstetrics and Gynecology, University of Cincinnati, Cincinnati, OH (TLP and TJ)
2 Supported by grants from the Swedish Research Council (grants 10838 and 14555), the Swedish Nutrition Foundation, the Swedish Diabetes Association, the Frimurare-Barnhus Direktionen, the Magnus Bergvall Foundation, the Åhlens Foundation, and the Wilhelm and Martina Lundgren Foundation. 3 Reprints not available. Address correspondence to N Jansson, Perinatal Center, Institute of Neurophysiology and Physiology, Department of Physiology, Göteborg University, PO Box 432, 405 30 Göteborg, Sweden. E-mail: nina.jansson{at}fysiologi.gu.se.
| ABSTRACT |
|---|
|
|
|---|
Objective: We aimed to identify maternal hormones that can explain the link between dietary intake, body mass index (BMI), and birth weight.
Design: Pregnant women with BMIs (in kg/m2) ranging from 17 to 44 (n = 49) were recruited in gestational weeks 8–12. Serum hormone concentrations were measured and dietary history interviews were performed in the first and third trimesters. Multiple regression models were produced to identify hormones that correlate with birth weight and are influenced by BMI or dietary factors.
Results: We found a strong positive correlation between BMI and first- and third-trimester insulin and leptin concentrations and a negative correlation between BMI and first-trimester adiponectin and first- and third-trimester insulin-like growth factor binding protein-1 (IGFBP-1). Maternal total fat intake in the first trimester was positively correlated with maternal leptin and inversely correlated with adiponectin. In addition, third-trimester total fat intake was positively correlated with circulating resistin concentrations. First-trimester maternal serum resistin was positively correlated with birth weight, whereas third-trimester maternal IGFBP-1 was negatively correlated with birth weight.
Conclusions: High first-trimester maternal serum resistin and low third-trimester IGFBP-1 were correlated with increased birth weight. We propose that low serum concentrations of IGFBP-1 represent a link between high BMI and increased fetal growth by increasing the bioavailability of insulin-like growth factor-I, which up-regulates placental nutrient transport.
| INTRODUCTION |
|---|
|
|
|---|
The mechanisms underlying fetal overgrowth in pregnancies complicated by maternal obesity are not known. Fetal growth is largely determined by nutrient transfer across the placenta, which is dependent on several factors, including maternal nutrient levels and placental transport capacity. Maternal hormones such as insulin (13, 14), leptin (14), and insulin-like growth factor-I (IGF-I) (15) have been shown to stimulate placental nutrient transporters (16). Leptin and fasting insulin are elevated in late pregnancy in obese women (17), and Clausen et al (18) reported that insulin concentrations in the second trimester were positively correlated with maternal first-trimester body mass index (BMI). Thus, it is possible that increased circulating concentrations of maternal hormones, such as leptin and insulin, constitute a mechanistic link between maternal overweight or obesity and fetal overgrowth mediated by up-regulation of placental nutrient transporters. Furthermore, obesity is associated with elevated maternal plasma concentrations of interleukin-6 (17) and decreased adiponectin concentrations (19) in late pregnancy.
It is well established that severe malnutrition in human pregnancy causes a reduction in birth weight (20). Also, more moderate variations in maternal diet in pregnancy can influence fetal growth (21–23). However, whether these effects are caused by altered maternal nutrient levels per se or by changes in plasma concentrations of maternal metabolic hormones is unknown. Indeed, animal experiments suggest that regulation of placental transport capacity by maternal hormones best explains altered fetal growth in response to changes in maternal dietary intake (24). Whether an increased dietary intake contributes to fetal overgrowth in women with overweight or obesity is not well established.
The previously demonstrated positive correlation between BMI and fetal growth (2, 5, 9, 25) provides no mechanistic information. The objective of the current study was to identify hormonal factors that can explain the link between early pregnancy BMI, maternal dietary intake, and birth weight. We focused on dietary factors reported in pregnant women or animal experiments to influence fetal growth and maternal hormones known to regulate placental nutrient transport in vitro.
| SUBJECTS AND METHODS |
|---|
|
|
|---|
Recruitment
Pregnant women of various adiposity as defined by their early pregnancy BMI [weight (kg)/height (m)2] were recruited at their first antenatal midwife visit (8–12 wk of gestation) in Göteborg, Sweden, between November 2003 and February 2006. The inclusion criteria were Scandinavian heritage, healthy, and
20 y old. The exclusion criteria were smoking, vegetarianism, assisted reproduction, concurrent disease such as eating disorder or diabetes, and development of pregnancy complications such as gestational diabetes, preeclampsia, or intrauterine growth restriction. Of the 56 pregnant women who were recruited, 7 were excluded during the course of the study because of either complications in their pregnancy (one subject each with miscarriage, gestational diabetes, stillbirth, or intrauterine growth restriction), moving away from the area (2 subjects), or lack of compliance with the study protocol after the first trimester (one subject). The remaining 49 women were healthy for the entire pregnancy and delivered between 36 and 43 wk of gestation.
Analyses of maternal serum hormones and adipokines
Maternal fasting blood samples were obtained at 2 time points during pregnancy: at recruitment (gestational weeks 8–12) and in the third trimester (weeks 32–35). Maternal venous samples were obtained in the morning by the midwives at the prenatal care center after the women had fasted overnight (no food after midnight). Blood samples were collected in tubes with and without anticoagulants (lithium heparin) and after 30 min were centrifuged for 10 min at 3000 rpm. Subsequently, tubes were sent to the Laboratory of Clinical Chemistry at Sahlgrenska University Hospital, Göteborg, Sweden, where serum and plasma were collected, aliquoted, and either analyzed directly or frozen at –80°C. Serum or plasma samples were analyzed for insulin (using immunometric methods), cortisol (immunochemistry), leptin and IGF-I (radioimmunoassay), and IGF binding protein-1 (IGFBP-1), adiponectin, and resistin (enzyme-linked immunosorbent assay). CVs (interassay precision) for these assays ranged from 3% to 12%.
Dietary assessment
Dietary intake data from all subjects were collected by a registered dietitian who interviewed the women at 8–12 wk and at 32–35 wk of gestation, respectively, by using the diet history method (26). The women were interviewed in person at the time of blood sampling after a light breakfast provided at the clinic. The interview was an open, structured, detailed interview designed to capture current habitual dietary intake, meal composition, and eating patterns.
The interview covered all dietary intake over a 24-h period, following the habitual dietary intake meal by meal, seeking day-to-day variation in food choice and meal composition, for weekdays and weekends. All women were asked about food and meal preparation methods, meals at restaurants, and fast food. The number of main-meals, in-between meals, snack-meals, night-meals, and nibbling were recorded. Information on quantities and portion sizes were collected at the end of the interview. Portion sizes of meals were described aided by using a food portion visual (Matmallen, Swedish National Food Administration, 1997). As all pregnant women, our study subjects obtained general dietary information in early pregnancy, at the first or second prenatal visit. However, none of the women in our cohort received special dietary advice from either the midwives or the dietitian. Dietary analyses were performed with the software ALTO DIET 32 (Aivo AB, Solna, Diet32 version 1.2.2.4) and were based on data from the Swedish National Food Administration. An average daily food intake was calculated by determining the mean of the 2 weekdays and multiplying that by 5/7 (5 weekdays out of 7 d) and adding that to the weekend day that was multiplied by 2/7 (2 weekend days out of 7 d).
Data collection
Anthropometric and medical data such as birth size (weight, length), placenta weight, sex, maternal height, weight gain, and pathologic conditions developing during pregnancy were obtained from medical charts after delivery.
Data presentation and statistical analysis
Descriptive univariate statistics including the mean, SE of the mean, and ranges were used to describe dietary intakes and hormone levels in our cohort. Differences between the first and the third trimester were analyzed by using Wilcoxon's signed-rank sum test for the hormones and the paired t test for dietary intake. To normalize birth weight and make it independent of sex and gestational age, z scores (standard scores) were calculated for all birth weights by using intrauterine growth curves for the Swedish population based on ultrasonically estimated fetal weights (27). The z score was derived by subtracting the population mean birth weight for sex and gestational age from an individual birth weight and then dividing the difference by the population SD. Relations between BMI and dietary factors, BMI and maternal hormones, dietary factors and hormones, and hormones and birth weight z scores, respectively, were determined by using bivariate analysis and either Pearson's or Spearman's correlation coefficients as appropriate. Subsequently, multiple regression explanatory models were produced for the first and third trimesters, respectively. First, BMI and dietary variables were used as explanatory variables for maternal serum hormone concentrations. Second, the relation between maternal hormones and birth weight z scores was analyzed. Analyses were carried out by using SPSS 11.0.4.
| RESULTS |
|---|
|
|
|---|
|
|
We found that both first- and third-trimester dietary energy intake was highly correlated with BMI (data not shown). In addition, BMI in the first trimester was significantly correlated with macronutrient intake (protein, total fat, saturated fat, monounsaturated fat, polyunsaturated fat, and carbohydrate intake). Similarly, maternal metabolic hormones and dietary variables were highly correlated. For example, insulin was positively and IGFBP-1 and adiponectin were negatively correlated with saturated and monounsaturated fat intake (data not shown).
Multiple regression explanatory models
To identify maternal hormones that could potentially mediate the effect of BMI and dietary factors on birth weight, we carried out multiple regression analyses in 2 steps. First, first-trimester BMI and dietary variables were used as explanatory variables for maternal serum hormone concentrations. Of the dietary variables, we chose to use 2 of the 3 main macronutrient classes in our analysis. Total fat was included because of highly significant correlations with most of the hormones in the bivariate analysis, and total protein intake was included because of reports in the literature of associations between protein intake and fetal growth (2, 28, 29). In this multiple regression model, we found a strong positive correlation between BMI and first- and third-trimester insulin and leptin and a negative correlation between BMI and first-trimester adiponectin and first- and third-trimester IGFBP-1 (Table 3
). With the exception of a positive correlation between first-trimester protein intake and leptin concentrations, protein intake had no effect on maternal hormone concentrations independent of BMI and total fat intake (Table 3
). Maternal intake of total fat in the first trimester was positively correlated with maternal serum leptin and inversely correlated with adiponectin concentrations. Interestingly, we found that third-trimester total fat intake was positively correlated with circulating resistin concentrations (Table 3
).
|
|
| DISCUSSION |
|---|
|
|
|---|
|
As evidenced by the positive correlations between early pregnancy BMI and the intake of protein, carbohydrate, and fat, obese women have a higher intake of all macronutrients, in particular in early pregnancy. These findings support the concept of obesity in pregnancy as a condition of overnutrition. There is a growing body of evidence from both animal and epidemiologic studies that maternal nutritional status affects fetal growth and development. However, most previous studies focused on the effects of maternal undernutrition (20, 34, 35, 36). Studies in women primarily focused on subjects within the normal BMI range and the relation between dietary intake and fetal growth. In some of these studies, no significant correlation between maternal energy or macronutrient intake and birth weight was shown (37, 38), whereas other studies reported that birth weight is associated with the percentage of energy obtained from dairy products in gestational week 16 (29), high carbohydrate intake in early pregnancy, and low intake of meat protein in relation to carbohydrate intake in late pregnancy (28). A significant effect of maternal protein intake on birth weight was also observed in a study by Cucó et al (23), who reported that proteins were the macronutrient with the largest effect on birth weight independent of energy intake or preconceptional BMI. The mechanisms linking dietary intake in pregnant women to fetal growth in these studies, however, are unclear.
In our model (Figure 1
), we proposed that one important mechanism by which diet influences fetal growth is by altering the circulating concentrations of key maternal metabolic hormones, which regulates placental nutrient transport and therefore fetal growth. This hypothesis was indirectly supported by significant correlations between dietary variables, in particular fat intake, and maternal serum concentrations of metabolic hormones. We therefore proceeded by carrying out a multiple regression analysis, which showed that first-trimester protein and fat intake were positively correlated with leptin concentrations and first-trimester fat intake was inversely associated with circulating adiponectin, associations that could not be explained by BMI. In the third trimester, total fat intake was correlated with serum resistin independently of BMI. These data show that maternal dietary intake variables influence concentrations of maternal hormones, which may alter fetal growth by affecting maternal metabolite levels and placental function.
We proceeded by assessing the relation between maternal hormones and birth weight. Using multiple regression modeling, we found that first-trimester maternal plasma resistin was positively and third-trimester maternal plasma IGFBP-1was negatively correlated with birth weight z score. IGFBP-1 generally inhibits IGF-I action and also has direct effects independent of IGF-I and the IGF-I receptor (39). We propose that the mechanistic link between low IGFBP-1 and increased fetal growth is increased IGF-I bioavailability. Ample experimental evidence exists showing that maternal IGF-I increases fetal growth and up-regulates placental nutrient transporters. Maternal IGF-I administered to pregnant guinea pigs midgestation increased placental and fetal growth (40). IGF-I has been shown to stimulate system A uptake in cultured term primary trophoblast cells (15) and to increase glucose and amino acid uptake in a first-trimester trophoblast cell line (41). Furthermore, in pregnant guinea pigs, in vivo evidence suggested that chronic infusion of IGF-I midgestation was associated with an increased placental capacity to transport glucose (42). Because BMI was inversely correlated with serum IGFBP-1 in our multiple regression analysis (Table 3
), serum IGFBP-1 concentrations may constitute one mechanistic link between BMI and birth weight.
Resistin, a cytokine secreted by adipocytes, was identified in 2001 and has been shown to cause insulin resistance in mice (43); however, the effects of resistin on insulin sensitivity in humans remain less well defined (44, 45). Maternal circulating concentrations of resistin are higher in pregnant than in nonpregnant women and have been reported to decrease with advancing gestation in women with normal BMI (46). In contrast, we observed an increase in maternal serum resistin between the first and third trimesters (Table 2
). This apparent discrepancy may be due to the inclusion of overweight and obese women in our cohort, because pregnant women with a high BMI have a high fat intake, and total fat intake in the third trimester was positively correlated with resistin concentrations (Table 3
). First-trimester resistin was positively correlated with birth weight, and this association was independent of BMI and dietary variables. The mechanism linking first-trimester resistin to fetal growth remains to be established. To the best of our knowledge, the effects of resistin on placental function and fetal growth have not been studied. However, it was recently reported that resistin stimulates lipoprotein lipase in 3T3-L1 adipocytes (47). If resistin has similar effects in placenta, it will result in increased transplacental transfer of free fatty acids, which may stimulate fetal growth and fat deposition.
In summary, high first-trimester maternal serum resistin and low third-trimester IGFBP-1 were correlated with increased birth weight. We propose that low serum concentrations of IGFBP-1 represent a link between high BMI and increased fetal growth by increasing the bioavailability of IGF-I, which up-regulates placental nutrient transport.
| ACKNOWLEDGMENTS |
|---|
The responsibilities of the authors were as follows—NJ: planning the study, patient recruitment, coordinator, data analysis and interpretation, and writing manuscript; AN: dietary interviews and data analysis; MG: statistical analyses, data interpretation, and contributing to writing the paper; MW: research funding and contributing to the planning of the study; LR-H: planning the study, interpreting the data, and contributing to the manuscript; TLP and TJ: planning the study, research funding, interpreting the data, and contributing to the manuscript. None of the authors had any conflicts of interest.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
H. N. Jones, L. A. Woollett, N. Barbour, P. D. Prasad, T. L. Powell, and T. Jansson High-fat diet before and during pregnancy causes marked up-regulation of placental nutrient transport and fetal overgrowth in C57/BL6 mice FASEB J, January 1, 2009; 23(1): 271 - 278. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |