Skip to main content
The Journal of Clinical Endocrinology and Metabolism logoLink to The Journal of Clinical Endocrinology and Metabolism
. 2015 Jan 9;100(4):1672–1680. doi: 10.1210/jc.2014-2949

Maternal Fuels and Metabolic Measures During Pregnancy and Neonatal Body Composition: The Healthy Start Study

Tessa L Crume 1,, Allison L Shapiro 1, John T Brinton 1, Deborah H Glueck 1, Mercedes Martinez 1, Mary Kohn 1, Curtis Harrod 1, Jacob E Friedman 1, Dana Dabelea 1
PMCID: PMC4399301  PMID: 25574704

Abstract

Context:

The impact of specific maternal fuels and metabolic measures during early and late gestation on neonatal body composition is not well defined.

Objective:

To determine how circulating maternal glucose, lipids, and insulin resistance in the first and second halves of pregnancy influence neonatal body composition.

Design:

A prospective pre-birth cohort enrolling pregnant women, the Healthy Start Study, was conducted, in which fasting maternal serum samples were collected twice during pregnancy to measure glucose, insulin, hemoglobin A1c, triglyceride, total cholesterol, high-density lipoprotein, and free fatty acids. Neonatal body composition was measured with air displacement plethysmography.

Setting:

An observational epidemiology study of pregnant women attending obstetric clinics at the University of Colorado, Anschutz Medical Center.

Participants:

This analysis includes 804 maternal-neonate pairs.

Results:

A strong positive linear relationship between maternal estimated insulin resistance (homeostasis model of assessment for insulin resistance) in the first half of pregnancy and neonatal fat mass (FM) and FM percentage (FM%) was detected, independent of prepregnancy body mass index (BMI). In the second half of pregnancy, positive linear relationships between maternal glucose levels and offspring FM and FM% were observed, independent of prepregnancy BMI. An inverse relationship was detected between high-density lipoprotein in the first half of pregnancy and FM, independent of prepregnancy BMI. Free fatty acid levels in the second half of pregnancy were positively associated with higher birth weight, independent of prepregnancy BMI.

Conclusion:

Maternal insulin resistance in the first half of pregnancy is highly predictive of neonatal FM%, whereas maternal glycemia, even within the normal range, is an important driver of neonatal adiposity in later pregnancy, independent of prepregnancy BMI. Our data provide additional insights on potential maternal factors responsible for fetal fat accretion and early development of adiposity.


Epidemiological data linking birth weight (as a marker of fetal growth) and various intrauterine exposures with an increased risk of adult diseases including heart disease, hypertension, and type 2 diabetes (13) have prompted interest in studying factors that influence fetal growth. Maternal weight and glucose homeostasis have been thought to be two of the most important determinants of fetal growth (4, 5). The association between maternal glycemia and increased birth weight has long been documented in diabetic pregnancies (6). However, macrosomia is not uncommon in diabetic pregnancies with strict glycemic control (7, 8) and among offspring of obese women with normal glucose tolerance (4, 9, 10), suggesting that glucose levels within the normal range and other fuels may contribute to fetal growth. Results from the Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study (11) made apparent the continuous relationship between maternal glucose levels measured in the third trimester, within a nondiabetic range, and a variety of maternal and neonatal outcomes, including birth weight, but also a surrogate measure of neonatal body fat based on a formula incorporating weight, length, and flank skinfold (12). Much less is known about the role of circulating maternal lipids on specific neonatal body compartments or the role of timing in gestation. A progressive elevation of circulating maternal lipids is observed in later gestation as a normal physiological response to optimize availability of substrates for fetal growth and development (13). Both maternal triglycerides (TGs) and free fatty acids (FFAs) in circulation have been correlated with neonatal weight and fat mass (FM) in women with gestational diabetes (14), suggesting that in this population, maternal hyperglycemia could influence lipid transfer to the fetus or that elevated lipid levels may alter glucose transfer pathways in the placenta (15). In some studies, maternal TG levels in mid to late pregnancy have been associated with increased birth weight after adjusting for prepregnancy body mass index (BMI) (1618). Several prospective studies have also shown an association between maternal circulating lipid levels (mostly TGs and high-density lipoprotein [HDL-c]) and increasing birth weight among women with hyperglycemia (17, 19, 20).

However, it remains unclear which maternal fuels and metabolic measures have the most substantial impact on fetal growth, body composition, and fat accretion and at which stage of pregnancy. The purpose of this report is to evaluate the association between levels of circulating maternal fuels (glucose and lipids) and metabolic measures (estimated insulin resistance) in two periods of gestation (first and second halves of pregnancy) with neonatal body size and composition, including FM, fat free mass (FFM), and FM percentage (FM%). We explored this association among 804 mother-infant pairs in the diverse pre-birth, prospective Healthy Start cohort study.

Subjects and Methods

Healthy Start Study

The Healthy Start Study recruited and enrolled pregnant women ≤24 weeks gestation from prenatal obstetrics clinics at University of Colorado Hospital in Aurora, Colorado. Enrollment criteria included women aged ≥ 16 years of age expecting a singleton birth, living in Colorado, and planning to deliver at University of Colorado Hospital. Women with serious chronic diseases (cancer, psychiatric diseases, steroid-dependent asthma, pre-existent diabetes), as well as those who subsequently experienced a fetal death or delivered a severely premature infant (<32 wk gestation) were excluded. Pregnant women who enrolled in the study were invited to participate in two in-person research visits, the first in early pregnancy (median gestational age, 17 wk; range, 11–20) and the second in mid to late pregnancy (median gestational age, 27 wk; range, 20–34). A third research visit was conducted in the hospital after delivery and before discharge (median, 1.0 d) when air displacement plethysmography (ADP) and anthropometric measurements were taken on neonates. All research measurements on mothers and neonates were obtained by trained clinical study nurses from the University of Colorado Hospital, Perinatal Clinical Translational Research Center. All participants received standard prenatal care from the obstetrics clinics at University of Colorado Hospital in Aurora, CO, including screening for gestational comorbidities.

Maternal measurements during pregnancy

At the two research visits during pregnancy, the following maternal information was collected: height measured with a stradiometer (Accustat; Genentech Inc); weight measured with a calibrated scale (Tanita); fasting venous blood samples; and questionnaires to assess demographic information, diet, and recent physical activity levels. Blood samples were drawn and analyzed at the University of Colorado Hospital Clinical and Translational Science Centers Core Laboratory for glucose, TGs, total cholesterol, HDL-c, FFA, and insulin levels. Insulin (μU/mL) was measured using a RIA by Millipore Corporation. Fasting glucose (mg/dL), cholesterol (mg/dL), HDL-c (mg/dL), TGs (mg/dL), and FFAs (mg/dL) were measured using manufacturer prepackaged enzymatic kits and the AU400e Chemistry Analyzer (Olympus America). Hemoglobin A1c (HbA1c) was measured at the second prenatal visit only from whole blood using a potassium ferricyanide assay by DCA Vantage Analyzer (Siemens). Maternal insulin resistance was estimated using the homeostasis model of assessment for insulin resistance (HOMA-IR) equation: glucose (mg/dL) * insulin (μU/mL)/405 (21).

Physical activity was assessed with the Pregnancy Physical Activity Questionnaire (22), a semiquantitative instrument that measures frequency and duration of time spent in four domains of activity: household/caregiving, occupational, sports/exercise, and transportation activities. Activities were assigned metabolic equivalent (MET) values according to the compendium of physical activities (23), and where possible, pregnancy-specific MET values (24). Reported duration of activity was multiplied by the respective MET value to estimate total energy expenditure (MET) in hours per week. Diet during pregnancy was collected with repeated 24-hour dietary recalls (up to six) using the Automated Self-Administered 24-Hour Dietary Recall (ASA24) of the National Cancer Institute (NCI) (25). The 24-hour recalls use automated computer technology, including graphic enhancements, animated characters to guide participants, and audio language/cues to enhance use in low-literacy populations as well as pictures of foods in multiple portion sizes to help respondents estimate portion size. In addition, two food propensity questionnaires were administered at the second prenatal visit and at delivery to capture usual intakes for particular foods or food groups over a 3-month period using an abbreviated food frequency instrument developed for the National Health and Nutrition Examination Survey, the “Dietary Screener” module (25).

Maternal prepregnancy BMI was calculated based on maternal prepregnancy weight obtained from medical records (89% of cases) or self-report (11% of cases), and maternal height was obtained at the first research visit. Gestational weight gain was estimated using a mixed effects model to predict gestational weight gain over a standardized length of gestation of 39 weeks, based on all available pregnancy weight measurements as well as maternal age, race/ethnicity, height, and BMI before pregnancy. The median number of weight measures per participant was 13, and the maximum number was 23. Prenatal smoking was ascertained through interview-administered questionnaires at each research visit. Clinically diagnosed gestational diabetes mellitus (GDM) status was determined based on a review of the maternal prenatal medical record.

Neonatal measures

Birth weight was measured using a calibrated scale, and birth length was measured to the nearest 0.1 cm using an infant board with a stadiometer. Other anthropometric factors as well as neonatal body composition were measured by trained nurses within 48 hours of birth (mean age, 1.5 ± 1.8 d). Neonatal body composition was measured by PEA POD (Life Measurement, Inc.), a two-compartment model measuring total FM (ie, adipose tissue) and FFM (ie, water, bone and non-bone mineral, and protein) in both absolute and proportionate terms using densitometric techniques based on ADP (26). The mean of the two closest measures out of a total of three obtained on each participant was used for each outcome. A standardized approach was used to estimate date of conception for each woman, and thus to derive gestational age. Each woman typically had multiple estimates of gestational age, all based on medical record abstractions: prenatal visits earlier in the pregnancy estimated gestational age from last menstrual period, whereas later prenatal visits estimated gestational age from ultrasound (available for 82% of participants). For each record of gestational age, an average conception date was derived for each woman, which was used to derive gestational age. Neonatal chronological age at PEA POD measure was calculated by taking the difference between the date of birth and the PEA POD research visit.

During recruitment, all mothers provided written informed consent. The Healthy Start Study protocol and procedures were approved by the Colorado Multiple Institutional Review Board.

Statistical analysis

Data in tables are presented as mean ± SD and number (percentage). The difference between maternal fuels/metabolic measures at the first and second visits was compared with a paired t test. Maternal dietary intake of macronutrients throughout pregnancy was estimated using mixed model regression analyses based on the NCI method (27), which adjusts for covariates (maternal age, day of the week) as well as variability of intake within and between individuals. Regression analyses were performed to determine the association of maternal metabolic fuels and metabolic measures (glucose, insulin, HbA1c, total cholesterol, HDL-c, FFA, and TG) measured at each visit with neonatal outcomes (birth weight, FM, FFM, and FM%). To assess the independent effect of maternal fuels/metabolic parameters in early vs mid to late pregnancy, we broke the correlation between the first and second visit fuel values by generating residual values representing the remaining variance in the fuel level at visit 1, adjusted for the fuel level at visit 2 and vice versa. A base model (model 1) adjusted for the residual value of the predictor from the other visit, infant sex, gestational age at birth, maternal age, race/ethnicity, and parity. To determine whether the relationship between each maternal fuel/metabolic measure and neonatal outcomes was modified by the gestational age at which the measurement was taken, interactions between each fuel measured at each visit and gestational age were evaluated in model 1. The interaction term was not significant at the preset significance level of ≤.10, thus it was dropped from the model. Maternal prepregnancy BMI was also assessed as a potential effect modifier of the relationship between maternal fuels/metabolic measures at each visit and neonatal outcomes in model 1, with the level of significance set at 0.10. Maternal prepregnancy BMI significantly modified the association between cholesterol levels during mid to late pregnancy and neonatal outcomes. For all other fuels, the interaction terms with maternal prepregnancy BMI were not significant. When no interaction was detected, further modeling proceeded as described below. Model 2 evaluated the confounding effect of factors during pregnancy, including average maternal daily calorie intake, average METs of daily physical activity, gestational weight gain, and smoking during pregnancy by adding these variables to model 1. Model 3 additionally adjusted for prepregnancy BMI if no effect modification was noted in model 1. Results are expressed as the effect of a 1-U change and SE in each maternal metabolic parameter on the neonatal outcomes.

Results

Maternal characteristics

A total of 1063 women were enrolled in the Healthy Start Study and delivered as of November 1, 2013, thus being eligible for this analysis. A total of 804 mother-neonate pairs (75%) had complete data on the variables of interest, including maternal fuels/metabolic measures, prepregnancy BMI, and neonatal PEA POD, and were included in the analysis. There were no relevant differences between the eligible and the analytic cohorts: mean maternal age, 27.7 vs 27.7 years; household income < $20,000/year, 18.8 vs 19.5%; maternal white non-Hispanic race, 53.8 vs 54.1%; mean maternal glucose level at the first prenatal visit, 76.7 vs 76.6 mg/dL; mean prepregnancy BMI, 25.6 vs 25.7 kg/m2; mean infant birth weight, 3222.8 vs 3250.6 g; mean neonatal FM, 291.5 vs 291.9 g, for the eligible and analytic cohorts, respectively.

The maternal and offspring characteristics are displayed in Table 1 and highlight the diversity of the cohort, with nearly half from race/ethnic groups other than white non-Hispanic, one-third with a level of education less than high school, 20% with an annual household income <$20,000 per year, and almost half with a prepregnancy BMI classified as overweight or obese. The mean gestational age at birth was 39.4 ± 1.3 weeks, and the mean birth weight was within the normal range, although slightly lower than that of the United States, likely due to the high altitude of Denver, Colorado (28).

Table 1.

Characteristics of Participating Mothers and Offspring at Birth

Mean ± SD or n (%)
Maternal characteristics
    Mean age, y 27.7 ± 6.1
    Race/ethnicity
        NHW 413 (53.6)
        Hispanic 184 (23.9)
        NHB 127 (16.5)
        Other 47 (6.1)
    Prepregnancy weight, kg 69.3 ± 17.5
    Height, cm 164.2 ± 6.9
    Prepregnancy BMI, kg/m2 25.7 ± 6.3
    Prepregnancy BMI status
        Underweight 26 (3.2)
        Healthy BMI 419 (52.1)
        Overweight 197 (24.5)
        Obese 162 (20.2)
    Household income <$20,000/y 124 (18.9)
    Maternal level of education < high school 245 (31.8)
    Primiparous 287 (35.8)
    Maternal smoking during pregnancy (any) 77 (9.6)
    Maternal GDM 26 (3.7)
Maternal diet and physical activity
    Total energy, kcal/d 2019.3 ± 550.8
    Total carbohydrates, g 254.2 ± 80.3
    Total fat, g 80.9 ± 25.1
    Calories from fat, % 34.2 ± 6.6
    Total saturated fat, g 27.5 ± 13.5
    Total physical activity (average METs) 194.6 ± 88.1
Offspring characteristics
    Gestational age, wk 39.4 ± 1.3
    Sex: female 418 (52.0)
    Birth weight, g 3246.6 ± 464.7
    Postnatal age at PEA POD, d 1.5 ± 1.8
    FM, g 290.1 ± 150.5
    FM, % 9.0 ± 3.9
    FFM, g 2822.2 ± 353.9
    Waist circumference, cm 29.4 ± 2.3
    Head circumference, cm 34.1 ± 2.2
    Sum of skinfolds, cm 15.1 ± 3.6

Abbreviations: NHW, non-Hispanic White; NHB, non-Hispanic Black.

Maternal fuels/metabolic parameters at each visit are also displayed in Table 2. Mean fasting plasma glucose, insulin, total cholesterol, HDL-c, and TG levels and HOMA-IR measures were significantly higher at the second vs the first pregnancy visit. Based on the International Association of Diabetes and Pregnancy Study Group Recommendations (29) for diagnosis of gestational diabetes, only 1.5% of mothers had a fasting glucose level higher than 92 mg/dL at the first visit and 3.1% of mothers at the second visit. HbA1c was measured at the late pregnancy visit only, and the mean level was 5.0 ± 0.3%. None of the subjects had an HbA1c level greater than 6.5%. With the exception of FFA levels, all lipid levels were significantly higher at the second pregnancy visit, with mean total cholesterol levels increasing by 15.1% and TG levels increasing by 30.4%.

Table 2.

Maternal Fuels/Metabolic Measures During First and Second Halves of Pregnancy

<20 wk Gestation ≥20 wk Gestation P Value
Maternal glucose, mg/dL 76.6 ± 6.8 77.8 ± 8.6 <.0001
Maternal insulin, μU/mL 14.1 ± 9.1 19.0 ± 15.9 <.0001
Fasting glucose >92 mg/dL, n (%) 12 (1.5) 22 (3.1) <.0001
HOMA-IR 2.7 ± 1.9 3.9 ± 0.3 <.0001
HbA1c, % 5.0 ± 0.3
Total cholesterol, mg/dL 182.3 ± 35.6 209.9 ± 40.3 <.0001
HDL-c, mg/dL 61.1 ± 12.6 63.1 ± 13.1 <.0001
FFAs, mg/dL 373.1 ± 166.0 365.1 ± 151.4 .3
TG, mg/dL 124.3 ± 49.6 162.2 ± 62.1 <.0001

Tables 3 and 4 present the results of multivariable analyses exploring the associations between maternal fuels/metabolic measures at each visit and neonatal outcomes: birth weight and FFM, Table 3; FM and FM%, Table 4.

Table 3.

Mean Difference ± SE in Neonatal Birth Weight and FFM per 1 U Increase in Maternal Fuels/Metabolic Measures in the First and Second Halves of Pregnancy

Birth Weight, g
FFM, g
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
β P β P β P β P β P β P
<20 wk gestation
    Fasting glucose, mg/dL 3.75 ± 1.96 .06 4.07 ± 2.10 .07 1.58 ± 2.25 .1 1.68 ± 1.53 .4 0.70 ± 1.58 .6 0.55 ± 1.50 .4
    HOMA-IR 25.50 ± 8.28 .002 23.90 ± 8.88 .007 5.39 ± 9.86 .6 6.01 ± 6.03 .3 8.50 ± 10.03 .4 8.48 ± 10.03 .4
    Total cholesterol, mg/dL 0.46 ± 0.39 .2 0.42 ± 0.42 .3 0.44 ± 0.41 .3 0.45 ± 0.28 .1 0.42 ± 0.30 .2 0.43 ± 0.30 .2
    HDL-c, mg/dL −0.54 ± 1.17 .6 −2.67 ± 1.22 .03 −1.71 ± 1.23 .2 0.26 ± 0.84 .8 −0.97 ± 0.89 .3 −0.57 ± 0.90 .5
    FFAs, mg/dL 0.06 ± 0.09 .5 0.05 ± 0.09 .6 −0.11 ± 0.10 .2 0.01 ± 0.06 .9 −0.01 ± 0.07 .9 −0.08 ± 0.07 .2
    TGs, mg/dL 0.09 ± 0.30 .7 0.50 ± 0.24 .04 0.41 ± 0.24 .08 0.09 ± 0.21 .7 0.34 ± 0.22 .1 0.25 ± 0.22 .3
≥20 wk gestation
    Fasting glucose, mg/dL 4.57 ± 1.69 .007 5.67 ± 1.80 .002 3.10 ± 1.88 .1 1.58 ± 1.19 .2 2.16 ± 1.27 .08 1.11 ± 1.32 .4
    HOMA-IR 3.50 ± 2.93 .2 4.96 ± 3.15 .02 1.29 ± 3.24 .7 0.60 ± 2.1 .7 1.40 ± 2.23 .5 −0.39 ± 2.31 .9
    HbA1c, % 122.0 ± 51.0 .01 114.96 ± 52.67 .03 142.75 ± 29.98 .2 43.46 ± 36.63 .2 48.57 ± 38.12 .2 31.73 ± 38.42 .4
    Total cholesterol, mg/dLa
    HDL-c, mg/dL −1.12 ± 1.12 .3 −3.12 ± 1.16 .007 −2.20 ± 1.16 .06 −0.60 ± 0.78 .5 −1.77 ± 0.83 .04 −1.39 ± 0.84 .1
    FFAs, mg/dL 0.21 ± 0.10 .03 0.31 ± 0.11 .003 0.24 ± 0.10 .02 0.09 ± 0.07 .2 0.16 ± 0.08 .04 0.12 ± 0.08 .2
    TGs, mg/dL 0.20 ± 0.24 .4 0.39 ± 0.24 .1 0.30 ± 0.24 .2 0.13 ± 0.17 .4 0.30 ± 0.17 .09 0.26 ± 0.17 .1

Model 1 is adjusted for the residual effect of the fuel level at the other visit, infant sex, gestational age, maternal age, race/ethnicity, parity, postnatal age at time of PEAPOD (for outcomes other than birth weight). Model 2 is model 1 plus maternal smoking, total energy intake, and maternal physical activity during pregnancy, gestational weight gain. Model 3 is model 2 plus prepregnancy BMI.

a

Effect modification detected with maternal prepregnancy BMI at the α = 0.10 level. See Figure 1 for stratified results.

Table 4.

Mean Difference ± SE in Neonatal FM and FM% per 1 U Increase in Maternal Fuels in the First and Second Halves of Pregnancy

FM, g
FM, %
Model 1
Model 2
Model 3
Model 1
Model 2
Model 3
β P β P β P β P β P β P
<20 wk gestation
    Fasting glucose, mg/dL 1.91 ± 0.81 .02 1.87 ± 0.85 .03 0.76 ± 0.86 .4 0.05 ± 0.02 .0004 0.05 ± 0.02 .03 0.02 ± 0.02 .3
    HOMA-IR 15.39 ± 3.18 <.0001 16.03 ± 3.41 <.0001 8.38 ± 3.78 .03 0.41 ± 0.08 <.0001 0.43 ± 0.09 <.0001 0.25 ± 0.10 .01
    Total cholesterol, mg/dL 0.01 ± 0.15 .9 −0.16 ± 0.16 .7 −0.05 ± 0.16 .7 −0.0008 ± 0.004 .8 −0.002 ± 0.004 .7 −0.002 ± 0.004 .6
    HDL-c, mg/dL −0.55 ± 0.45 .2 −1.41 ± 0.48 .003 −1.00 ± 0.47 .03 −0.02 ± 0.01 .2 −0.03 ± 0.01 .005 −0.2 ± 0.01 .05
    FFAs, mg/dL 0.01 ± 0.03 .7 0.02 ± 0.03 .6 −0.06 ± 0.04 .1 0.0003 ± 0.0008 .7 0.0005 ± 0.0009 .6 −0.001 ± 0.001 .1
    TGs, mg/dL 0.05 ± 0.11 .7 0.21 ± 0.12 .07 0.12 ± 0.11 .3 0.001 ± 0.003 .7 0.005 ± 0.003 .08 0.003 ± 0.003 .3
≥20 wk gestation
    Fasting glucose, mg/dL 2.74 ± 0.63 <.0001 3.06 ± 0.67 <.0001 1.97 ± 0.69 .004 0.07 ± 0.02 <.0001 0.08 ± 0.02 <.0001 0.05 ± 0.02 .004
    HOMA-IR 2.8 ± 1.1 .01 2.75 ± 1.17 .02 1.32 ± 1.20 .3 0.08 ± 0.03 .008 0.07 ± 0.03 .02 0.04 ± 0.03 .2
    HbA1c, % 69.8 ± 19.2 .0003 62.08 ± 19.92 .002 44.67 ± 19.74 .02 1.74 ± 0.50 .0006 1.50 ± 0.53 .005 1.06 ± 0.52 .04
    Total cholesterol, mg/dLa
    HDL-c, mg/dL −0.36 ± 0.43 .4 −1.16 ± 0.44 .01 −0.76 ± 0.44 .09 −0.006 ± 0.001 .6 −0.02 ± 0.01 .03 −0.01 ± 0.06 .2
    FFA, mg/dL 0.06 ± 0.04 .1 0.09 ± 0.04 .04 0.05 ± 0.04 .2 0.001 ± 0.0009 .2 0.002 ± 0.001 .09 0.0009 ± 0.001 .4
    TGs, mg/dL 0.05 ± 0.09 .6 0.16 ± 0.09 .09 0.12 ± 0.09 .2 0.001 ± 0.002 .6 0.004 ± 0.002 .1 0.003 ± 0.002 .2

Model 1 is adjusted for the residual effect of the fuel level at the other visit, infant sex, gestational age, maternal age, race/ethnicity, parity, postnatal age at time of PEA POD (for outcomes other than birth weight). Model 2 is model 1 plus maternal smoking, total energy intake, and maternal physical activity during pregnancy, gestational weight gain. Model 3 is model 2 plus prepregnancy BMI.

a

Effect modification detected with maternal prepregnancy BMI at the α = 0.10 level. See Figure 1 for stratified results.

First half of pregnancy findings

In model 1, each 1 mg/dL increase in fasting glucose levels was associated with a 1.9 g higher FM (P = .02) and a 0.05% increase in FM% (P = .0004). Each 1 U increase in HOMA-IR was associated with a 25.5 g higher birth weight (P = .002), 15.4 g higher FM (P < .0001), and a 0.41% increase in FM% (P < .0001). Adjustment for intermediates in model 2 (maternal smoking, total calorie intake, physical activity levels, and gestational weight gain) had little effect, and all the above relationships remained significant. In addition, in model 2, each 1 mg/dL increase in HDL-c was associated with a 2.67-g reduction in birth weight (P = .03), a 1.41-g reduction in FM (P = .003), and a 0.03% reduction in FM% (P = .005). Also in model 2, each 1 mg/dL increase in TG levels was associated with a 0.50-g increase in birth weight (P = .04). After adjustment for prepregnancy BMI in model 3, statistically significant relationships remained between HOMA-IR and FM (P = .03) as well as FM% (P = .01), and between HDL-c and FM (P = .03); and there was borderline significance with FM% (P = .05). There were no statistically significant relationships between total cholesterol or FFA and neonatal body size or composition.

Second half of pregnancy findings

In model 1, each 1 mg/dL increase in fasting glucose was associated with a 4.6 g higher birth weight (P = .007), a 2.7 g higher FM (P < .0001), and a 0.07% increase in FM% (P = .0004). HOMA-IR was associated with a 2.8 g higher FM (P = .01) and an increase in FM% of 0.08% (P = .008). Similarly, each 1% increase in HbA1c was associated with a 122.0 g higher birth weight (P = .01), a 69.8 g higher FM (P = .0003), and a 1.7% higher FM% (P = .0006). Each 1 mg/dL increase in FFA was associated with a 0.21 g increase in birth weight (P = .03). Upon adjustment for smoking, total calorie intake, physical activity, and gestational weight gain in model 2, relationships with glucose, HOMA-IR, and HbA1c remained; HDL-c was inversely associated with all neonatal outcomes; and FFAs were positively associated with birth weight (P = .003), FM (P = .04), and FFM (P = .04). After adjustment for prepregnancy BMI in model 3, only the following relationships remained significant: glucose and FM% (P = .004); HbA1c and FM (P = .02) as well as FM% (P = .04); FFA and birth weight (P = .02).

A significant interaction was detected between total cholesterol levels in the second half of pregnancy and maternal prepregnant BMI on the relationship with neonatal outcomes. This relationship was explored for a range of prepregnancy BMIs and is presented in Figure 1. Figure 1A shows a positive linear relationship between cholesterol levels and birth weight for higher prepregnancy BMIs and a null effect for lean women. In Figure 1B, the slope of cholesterol levels on FM was positive at higher prepregnancy BMI, null for healthy BMIs, and negative for BMI in the underweight range. Similar to birth weight, Figure 1C shows a positive association between cholesterol levels and FFM for women with higher prepregnancy BMIs and a null effect for lean women. Finally, in Figure 1D, there was a positive association between cholesterol levels and FM% at higher prepregnancy BMIs, a null effect for lean BMIs, and an inverse trend for BMIs in the underweight range.

Figure 1.

Figure 1.

Relationship between maternal cholesterol in mid/late pregnancy and neonatal body size and composition by prepregnancy BMI. A, Birth weight in grams; B, FM in grams; C, FFM in grams; D, FM%.

Our findings were not influenced by the exclusion of women identified with GDM (n = 26), gestational hypertension (n = 61), or pre-eclampsia (n = 34).

Discussion

In a diverse, contemporary cohort of mother-infant pairs, we found that increasing maternal fasting glucose levels and estimated insulin resistance (HOMA-IR) in both early and late pregnancy were linearly and positively associated with neonatal FM and FM%, independent of maternal age, parity, race/ethnicity, diet, physical activity, and gestational weight gain. Although the associations of neonatal adiposity parameters with glucose levels in the first half of pregnancy seem to be largely confounded by prepregnancy BMI, the relationships in late pregnancy remained after adjustment. Similarly, HbA1c levels were strongly associated with neonatal adiposity, independent of maternal BMI. In contrast, HOMA-IR in the first half of pregnancy had a strong influence on FM and FM%, independent of prepregnancy BMI. These results suggest time-dependent specific intrauterine effects of maternal insulin resistance and glucose levels during pregnancy on fetal and neonatal fat accretion.

Our findings of significant associations between maternal fasting glucose levels in late pregnancy, within a nondiabetic range, and higher birth weight and neonatal adiposity are consistent with findings reported by the HAPO study (10, 11). Although HAPO estimated neonatal adiposity based on a formula incorporating weight, length, and flank skinfold measurement (12), we used a direct measurement of neonatal total FM and FFM based on ADP. In addition, our findings suggest that maternal glucose levels in late pregnancy, adjusted for those in early pregnancy, are particularly important for neonatal fat accretion and that this association is not confounded or modified by maternal BMI. During the last third of gestation, due to the demands of the placental-fetal unit and rapid depletion of glycogen stores, the mother switches to a catabolic state in which glucose is the predominant nutrient crossing the placenta and maternal adipose tissue lipolytic activity is enhanced (30). Among pregnant women infused with glucose labeled with stable isotopes several hours before delivery, it was shown that 95% of infant plasma glucose after birth was from maternal plasma (31). This supports the notion that maternal glycemia in late pregnancy may play an important role in neonatal body fat accretion across the entire range of maternal glucose levels.

The mechanisms whereby increasing levels of maternal insulin resistance alter body composition in the fetus are unclear, but results of this study suggest that they may be independent of maternal prepregnancy obesity, at least during the first half of pregnancy. Maternal insulin does not cross the placenta; however, it may up-regulate nutrient sensors on the placenta, which in turn stimulate fetal cell proliferation and growth (32). Early gestation insulin resistance may affect placenta transport mechanisms and could contribute to factors such as increased leptin, adiponectin, or inflammation, thus indirectly influencing neonatal adiposity. Studies have demonstrated an up-regulation of amino acid transporter isoforms, activated by insulin, in the placentas of women with obesity who gave birth to large-for-gestational age babies (33), suggesting that maternal insulin may indirectly mediate fetal growth.

We found less consistent results on relationships of maternal lipid levels with birth weight and neonatal body composition. HDL-c levels throughout pregnancy were inversely associated with all neonatal body composition parameters, but only after adjustment for intermediate variables, including gestational weight gain. FFA levels during late gestation were positively associated with birth weight and FM, although only the former association was independent of pregnancy BMI. In addition, TG levels in late pregnancy were associated with birth weight but were not independent of prepregnancy BMI. Previous findings on the role of circulating maternal fuels other than glucose during pregnancy have been mixed. The Pune Maternal Nutrition Study (34) found that maternal glucose, total cholesterol, and TG levels at both 18 and 28 weeks gestation were associated with higher birth weight among a population of women with low prepregnancy BMI. However, they did not estimate effects on specific neonatal body composition compartments. Misra et al (35) assessed the relationship between maternal serum lipid levels and birth weight among 143 women-infant pairs from Michigan and reported an inverse relationship between HDL-c levels and birth weight among overweight/obese women, but not normal-weight women. The investigators also found that maternal TG levels were positively associated with birth weight, although only among normal-weight women. Similarly, in a nested case control study of risk factors associated with macrosomia in Oslo, Norway, Clausen et al (36) found that high serum insulin and HDL-c were associated with an increased risk of macrosomia. Finally, in a highly controlled small clinical study of normal-weight (n = 22) and obese (n = 16) pregnant women in Colorado on fixed diets, Harmon et al (37) reported that early pregnancy fasting TG and late pregnancy fasting FFAs were more strongly correlated with infant adiposity (based on skinfolds assessment) than late pregnancy glucose and fasting insulin. Schaefer-Graf et al (38) found no relationship between FFA levels and infant birth weight or FM among 190 mother-infant pairs with normal glucose tolerance in Germany.

Our data on a large diverse cohort found a relatively modest impact of maternal lipids on neonatal FM accretion in a general population of pregnant women, relative to the impact of glucose and insulin resistance. Studies in rodents have shown that glucose is quantitatively the most abundant of several substrates that cross the placenta (39). Enhanced adipose tissue lipolytic activity has been reported in pregnant women (40) and animal models (41). Maternal TGs are hydrolyzed to FFA to cross the placenta, and FFAs cross in small proportions (42); however, plasma ketone bodies easily cross the placenta and serve not only as fuels for the fetus, but also as lipogenic substrates (43). Insulin resistance is thought to enhance adipose tissue lipolytic activity (44) and decrease adipose tissue lipoprotein lipase activity (45); thus, it may mediate the relationship between maternal lipids and fetal FM accretion.

Another interesting finding in our study was the modification of effects of maternal cholesterol levels in late pregnancy on all neonatal body composition measures by prepregnancy BMI. A positive effect was noted for all neonatal outcomes at higher prepregnancy BMIs, with a null effect for lean women and an inverse relationship on FM for underweight women. During the last trimester of pregnancy, maternally derived cholesterol contributes 22–40% of the fetal cholesterol pool (46). Maternal cholesterol is available for fetal use to build cell membranes and as a precursor of bile acids, steroid hormones, and cell proliferation for the development of the growing body. Our findings suggest a joint effect of circulating maternal cholesterol levels and maternal BMI on FM and FFM development, which needs to be further explored in mechanistic studies.

Our study has some limitations and numerous strengths. The two study visits were conducted at a median gestational age of 17 weeks (range, 11–20) and 27 weeks (range, 20–34) of pregnancy, which, although spanning almost the entire duration of pregnancy, limits our ability to assess trimester-specific effects of maternal fuels/metabolic measures. Due to the dependence of neonatal body composition outcomes on gestational age at birth, we adjusted for this in our models. However, this could introduce bias because unmeasured pathologies can trigger both earlier delivery and fetal growth (ie, act as collider). To address this concern, models omitting gestational age were also explored, with no substantial influence on our findings. In addition, gestational age at birth was calculated based on date of conception, estimated by averaging the self-reported last day of menstrual period and ultrasound-based estimation. Conception date based on self-reported last day of menstrual period has been shown to be less accurate for mothers that are young, primiparous, smokers, or obese/overweight, whereas estimates based on ultrasound are biased higher for larger fetuses (47). Although there may be some degree of misclassification of gestational age, it is unlikely to have significantly biased our results because effect estimates were minimally impacted by omitting gestational age from our multivariable models. Although causal inference is limited with an observational study, the prospective and longitudinal design of Healthy Start provides strong evidence for temporality and order of events. Moreover, our study collected high-quality repeated measurements of maternal circulating metabolic parameters during pregnancy, standardized maternal and infant anthropometry data, as well as direct measures of neonatal body composition immediately after birth. Finally, this is a large cohort, representing a contemporary U.S. urban population with diverse racial/ethnic and socioeconomic composition.

Conclusions

We provide novel and comprehensive evidence for the role of various maternal fuels/metabolic measures during both early and late pregnancy on fetal growth, neonatal body size, and composition. Our data suggest that early and late gestation fuel switching may play different roles in mechanisms responsible for fetal fat accretion and development of adiposity. In the first half of pregnancy, maternal insulin resistance is an independent predictor of neonatal adiposity; later on, maternal glycemia, even within the normal range, becomes a main driver of fat accretion. Importantly, these relationships are independent of prepregnancy BMI and gestational weight gain. Maternal circulating lipid levels have a relatively modest impact on neonatal FM accretion and adiposity, although studies are necessary to further explore and replicate effect modification by maternal prepregnant BMI. Future mechanistic studies should further elucidate the various pathways responsible for these associations.

Acknowledgments

We gratefully acknowledge the mothers and babies that participated in the Healthy Start Study.

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK076648 (Principal Investigator, Dana Dabelea).

Clinical Trial Registration: NCT02273297.

Disclosure Summary: The authors have nothing to disclose.

Footnotes

Abbreviations:
ADP
air displacement plethysmography
BMI
body mass index
FFA
free fatty acid
FFM
fat free mass
FM
fat mass
GDM
gestational diabetes mellitus
HbA1c
hemoglobin A1c
HDL-c
high-density lipoprotein
HOMA-IR
homeostasis model of assessment for insulin resistance
MET
metabolic equivalent
TG
triglyceride.

References

  • 1. Barker DJ. In utero programming of cardiovascular disease. Theriogenology. 2000;53(2):555–574. [DOI] [PubMed] [Google Scholar]
  • 2. Barker DJ. The fetal origins of type 2 diabetes mellitus. Ann Intern Med. 1999;130:322–324. [DOI] [PubMed] [Google Scholar]
  • 3. Barker DJ. The fetal origins of hypertension. J Hypertens Suppl. 1996;14(5):S117–S120. [PubMed] [Google Scholar]
  • 4. Sewell MF, Huston-Presley L, Super DM, Catalano P. Increased neonatal fat mass, not lean body mass, is associated with maternal obesity. Am J Obstet Gynecol. 2006;195(4):1100–1103. [DOI] [PubMed] [Google Scholar]
  • 5. Durnwald C, Landon M. Fetal links to chronic disease: the role of gestational diabetes mellitus. Am J Perinatol. 2013;30(5):343–346. [DOI] [PubMed] [Google Scholar]
  • 6. Pedersen J. Glucose content of the amniotic fluid in diabetic pregnancies; correlations with the maternal blood sugar. Acta Endocrinol (Copenh). 1954;15(4):342–354. [DOI] [PubMed] [Google Scholar]
  • 7. Schwartz R, Gruppuso PA, Petzold K, Brambilla D, Hiilesmaa V, Teramo KA. Hyperinsulinemia and macrosomia in the fetus of the diabetic mother. Diabetes Care. 1994;17(7):640–648. [DOI] [PubMed] [Google Scholar]
  • 8. Evers IM, de Valk HW, Mol BW, ter Braak EW, Visser GH. Macrosomia despite good glycaemic control in type I diabetic pregnancy; results of a nationwide study in The Netherlands. Diabetologia. 2002;45(11):1484–1489. [DOI] [PubMed] [Google Scholar]
  • 9. Tallarigo L, Giampietro O, Penno G, Miccoli R, Gregori G, Navalesi R. Relation of glucose tolerance to complications of pregnancy in nondiabetic women. N Engl J Med. 1986;315(16):989–992. [DOI] [PubMed] [Google Scholar]
  • 10. HAPO Study Cooperative Research Group, Metzger BE, Lowe LP, Dyer AR, et al. Hyperglycemia and adverse pregnancy outcomes. N Engl J Med. 2008;358(19):1991–2002. [DOI] [PubMed] [Google Scholar]
  • 11. HAPO Study Cooperative Research Group. Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study: associations with neonatal anthropometrics. Diabetes. 2009;58:453–459. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Catalano PM, Thomas AJ, Avallone DA, Amini SB. Anthropometric estimation of neonatal body composition. Am J Obstet Gynecol. 1995;173(4):1176–1181. [DOI] [PubMed] [Google Scholar]
  • 13. Herrera E. Implications of dietary fatty acids during pregnancy on placental, fetal and postnatal development–a review. Placenta. 2002;23(suppl A):S9–S19. [DOI] [PubMed] [Google Scholar]
  • 14. Schaefer-Graf UM, Graf K, Kulbacka I, et al. Maternal lipids as strong determinants of fetal environment and growth in pregnancies with gestational diabetes mellitus. Diabetes Care. 2008;31(9):1858–1863. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Bibee KP, Illsley NP, Moley KH. Asymmetric syncytial expression of GLUT9 splice variants in human term placenta and alterations in diabetic pregnancies. Reprod Sci. 2011;18(1):20–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Di Cianni G, Miccoli R, Volpe L, Lencioni C, Ghio A, Giovannitti MG, Cuccuru I, Pellegrini G, Chatzianagnostou K, Boldrini A, Del Prato S. Maternal triglyceride levels and newborn weight in pregnant women with normal glucose tolerance. Diabet Med. 2005;22(1):21–25. [DOI] [PubMed] [Google Scholar]
  • 17. Kitajima M, Oka S, Yasuhi I, Fukuda M, Rii Y, Ishimaru T. Maternal serum triglyceride at 24–32 weeks' gestation and newborn weight in nondiabetic women with positive diabetic screens. Obstet Gynecol. 2001;97:776–780. [DOI] [PubMed] [Google Scholar]
  • 18. Nolan CJ, Riley SF, Sheedy MT, Walstab JE, Beischer NA. Maternal serum triglyceride, glucose tolerance, and neonatal birth weight ratio in pregnancy. Diabetes Care. 1995;18(12):1550–1556. [DOI] [PubMed] [Google Scholar]
  • 19. Emet T, Ustüner I, Güven SG, et al. Plasma lipids and lipoproteins during pregnancy and related pregnancy outcomes. Arch Gynecol Obstet. 2013;288(1):49–55. [DOI] [PubMed] [Google Scholar]
  • 20. Knopp RH, Walden CE, Wahl PW, et al. Effect of postpartum lactation on lipoprotein lipids and apoproteins. J Clin Endocrinol Metab. 1985;60(3):542–547. [DOI] [PubMed] [Google Scholar]
  • 21. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412–419. [DOI] [PubMed] [Google Scholar]
  • 22. Chasan-Taber L, Schmidt MD, Roberts DE, Hosmer D, Markenson G, Freedson PS. Development and validation of a Pregnancy Physical Activity Questionnaire. Med Sci Sports Exerc. 2004;36(10):1750–1760. [DOI] [PubMed] [Google Scholar]
  • 23. Ainsworth BE, Haskell WL, Whitt MC, et al. Compendium of physical activities: an update of activity codes and MET intensities. Med Sci Sports Exerc. 2000;32(9 suppl):S498–S504. [DOI] [PubMed] [Google Scholar]
  • 24. Chasan-Taber L, Freedson PS, Roberts DE, Schmidt MD, Fragala MS. Energy expenditure of selected household activities during pregnancy. Res Q Exerc Sport. 2007;78(2):133–137. [DOI] [PubMed] [Google Scholar]
  • 25. Subar AF, Kirkpatrick SI, Mittl B, et al. The Automated Self-Administered 24-hour dietary recall (ASA24): a resource for researchers, clinicians, and educators from the National Cancer Institute. J Acad Nutr Diet. 2012;112(8):1134–1137. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Ma G, Yao M, Liu Y, et al. Validation of a new pediatric air-displacement plethysmograph for assessing body composition in infants. Am J Clin Nutr. 2004;79(4):653–660. [DOI] [PubMed] [Google Scholar]
  • 27. Tooze JA, Kipnis V, Buckman DW, et al. A mixed-effects model approach for estimating the distribution of usual intake of nutrients: the NCI method. Stat Med. 2010;29(27):2857–2868. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Jensen GM, Moore LG. The effect of high altitude and other risk factors on birthweight: independent or interactive effects? Am J Public Health. 1997;87(6):1003–1007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. International Association of Diabetes and Pregnancy Study Groups Consensus Panel, Metzger BE, Gabbe SG, Persson B, Buchanan TA, Catalano PA, Damm P, Dyer AR, Leiva Ad, Hod M, Kitzmiler JL, Lowe LP, McIntyre HD, Oats JJ, Omori Y, Schmidt MI. International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care. 2010;33:676–682. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Herrera E. Metabolic adaptations in pregnancy and their implications for the availability of substrates to the fetus. Eur J Clin Nutr. 2000;54(suppl 1):S47–S51. [DOI] [PubMed] [Google Scholar]
  • 31. Staat BC, Galan HL, Harwwod JE, Lee G, Marconi AM, Paolini CL, Cheung A, Battaglia FC. Transplancental supply of mannose and inositol in uncomplicated pregnancies using stable isotopes. J Clin Endocrinol Metab. 2012; 97:2497–2502. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Jansson T, Aye IL, Goberdhan DC. The emerging role of mTORC1 signaling in placental nutrient-sensing. Placenta. 2012;33(suppl 2):e23–e29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Jansson N, Rosario FJ, Gaccioli F, et al. Activation of placental mTOR signaling and amino acid transporters in obese women giving birth to large babies. J Clin Endocrinol Metab. 2013;98(1):105–113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Kulkarni SR, Kumaran K, Rao SR, et al. Maternal lipids are as important as glucose for fetal growth: findings from the Pune Maternal Nutrition Study. Diabetes Care. 2013;36(9):2706–2713. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Misra VK, Trudeau S, Perni U. Maternal serum lipids during pregnancy and infant birth weight: the influence of prepregnancy BMI. Obesity (Silver Spring). 2011;19(7):1476–1481. [DOI] [PubMed] [Google Scholar]
  • 36. Clausen T, Burski TK, Øyen N, Godang K, Bollerslev J, Henriksen T. Maternal anthropometric and metabolic factors in the first half of pregnancy and risk of neonatal macrosomia in term pregnancies. A prospective study. Eur J Endocrinol. 2005;153(6):887–894. [DOI] [PubMed] [Google Scholar]
  • 37. Harmon KA, Gerard L, Jensen DR, et al. Continuous glucose profiles in obese and normal-weight pregnant women on a controlled diet: metabolic determinants of fetal growth. Diabetes Care. 2011;34(10):2198–2204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Schaefer-Graf UM, Meitzner K, Ortega-Senovilla H, et al. Differences in the implications of maternal lipids on fetal metabolism and growth between gestational diabetes mellitus and control pregnancies. Diabet Med. 2011;28(9):1053–1059. [DOI] [PubMed] [Google Scholar]
  • 39. Lasunción MA, Lorenzo J, Palacin M, Herrera E. Maternal factors modulating nutrient transfer to fetus. Biol Neonate. 1987;51(2):86–93. [DOI] [PubMed] [Google Scholar]
  • 40. Williams C, Coltart TM. Adipose tissue metabolism in pregnancy: the lipolytic effect of human placental lactogen. Br J Obstet Gynaecol. 1978;85(1):43–46. [DOI] [PubMed] [Google Scholar]
  • 41. Knopp RH, Herrera E, Freinkel N. Carbohydrate metabolism in pregnancy. 8. Metabolism of adipose tissue isolated from fed and fasted pregnant rats during late gestation. J Clin Invest. 1970;49(7):1438–1446. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Herrera E, Amusquivar E, López-Soldado I, Ortega H. Maternal lipid metabolism and placental lipid transfer. Horm Res. 2006;65(suppl 3):59–64. [DOI] [PubMed] [Google Scholar]
  • 43. Shambaugh GE. Ketone body metabolism in the mother and fetus. Fed Proc. 1985; 44(7):2347–2351. [PubMed] [Google Scholar]
  • 44. Ramos P, Herrera E. Reversion of insulin resistance in the rat during late pregnancy by 72-h glucose infusion. Am J Physiol. 1995;269:E858–E863. [DOI] [PubMed] [Google Scholar]
  • 45. Knopp RH, Warth MR, Charles D, et al. Lipoprotein metabolism in pregnancy, fat transport to the fetus, and the effects of diabetes. Biol Neonate. 1986;50(6):297–317. [DOI] [PubMed] [Google Scholar]
  • 46. Pitkin RM, Connor WE, Lin DS. Cholesterol metabolism and placental transfer in the pregnant rhesus monkey. J Clin Invest. 1972;51(10):2584–2592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Lynch CD, Zhang J. The research implications of the selection of a gestational age estimation method. Paediatr Perinat Epidemiol. 2007;21(suppl 2):86–96. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Clinical Endocrinology and Metabolism are provided here courtesy of The Endocrine Society

RESOURCES