Abstract
Background
In adults, adherence to the Mediterranean diet has been inversely associated with cardiovascular risk, but the extent to which diet in pregnancy is associated with offspring adiposity is unclear. We aimed to investigate the association between adherence to Mediterranean diet in pregnancy and offspring cardiometabolic traits in two pregnancy cohorts.
Methods
We studied 997 mother-child pairs from Project Viva in Massachusetts, USA, and 569 pairs from the Rhea study in Crete, Greece. We estimated adherence to the Mediterranean diet with an a priori defined score (MDS) of 9 foods and nutrients (0 to 9). We measured child weight, height, waist circumference, skin fold thicknesses, blood pressure (BP), and blood levels of lipids, c-reactive protein, and adipokines in mid-childhood (median 7.7 years) in Viva, and in early childhood (median 4.2 years) in Rhea. We calculated cohort-specific effects, and pooled effects estimates with random-effects models for cohort and child age.
Results
In Project Viva the mean (SD) MDS was 2.7 (1.6); in Rhea it was 3.8 (1.7). In the pooled analysis, for each 3-point increment in the MDS, offspring BMI z score was lower by 0.14 units (95% CI, −0.15 to −0.13), waist circumference by 0.39cm (95% CI, −0.64 to −0.14), and the sum of skin fold thicknesses by 0.63mm (95% CI, −0.98 to −0.28). We also observed lower offspring systolic (−1.03 mmHg; 95% CI, −1.65 to −0.42) and diastolic BP (−0.57mmHg; 95% CI, −0.98 to −0.16).
Conclusion
Greater adherence to Mediterranean diet during pregnancy may protect against excess offspring cardiometabolic risk.
Keywords: diet, pregnancy, cohort study, obesity, blood pressure, lipids
INTRODUCTION
Early life is a critical period of developmental plasticity [1]. Metabolic programming is the phenomenon whereby a nutritional stress/stimulus applied during critical periods of early development permanently alters an organism’s physiology and metabolism, the consequences of which are often observed much later in life [2]. Although the fetal origins hypothesis, proposed by Barker, has been well documented in animal studies, data from human studies on maternal diet quality during pregnancy in association with offspring cardiometabolic risk factors are scarce with disparate results [3,4]. Prior studies have examined associations of specific nutrients, foods, or food groups during pregnancy with offspring health, and these approaches may not take into consideration that some nutrients are interrelated. Moreover, they all refer to a specific population group and their results cannot be easily generalized due to cultural and socioeconomic population differences associated with diet.
The traditional Mediterranean diet is characterized by a high intake of olive oil, fruits, vegetables, legumes, nuts, and whole grain products; a moderate intake of fish; and only small amounts of red and processed meat. This dietary pattern is low in saturated fat intake and high in monounsaturated fat intake from olive oil, it is rich in fibre and glutathione, provides a balanced ratio of n-6/n-3 essential fatty acids, and high amounts of antioxidants (especially polyphenols from olive oil, vitamins E and C) [5]. Several epidemiological studies and clinical trials support the role of the Mediterranean diet in preventing obesity, type 2 diabetes mellitus and metabolic syndrome in adults [6,7], while some recent studies suggest a protective role against obesity development in children [8,9]. In pregnancy, a higher adherence to the Mediterranean diet has been associated with lower risk of preterm birth [10,11], higher birth weight [12,13] and lower offspring waist circumference at preschool age [14].
The objective of this study was to investigate associations of maternal adherence to the Mediterranean diet in early pregnancy with offspring obesity and cardiometabolic risk in two cohorts with different socioeconomic characteristics and different geographic locations: Project Viva, a prospective mother-child cohort that was established in Massachusetts, USA in 1999 [15] and the Rhea study, a population based mother-child cohort initiated in Crete, Greece in 2007 [16].
SUBJECTS AND METHODS
Study population
Project Viva
We recruited women at their first prenatal visit from Atrius Harvard Vanguard Medical Associates, a multi-specialty group practice in Massachusetts, from 1999 to 2003 [15]. Of 2128 live singleton births in Project Viva, 1784 pregnant women provided information on first trimester diet via a validated semi-quantitative food frequency questionnaire [17]. We excluded 7 women with implausible values of energy intake (<600 or >6000 calories/day). Of the 1777 remaining mother-child pairs, 997 attended an in-person mid-childhood visit at 6–10 years (median 7.7 years; IQR: 7.3–8.3), during which study staff measured anthropometry and collected fasting blood samples.
The Rhea cohort
The Rhea project is a population-based cohort of pregnant women and their children in the prefecture of Heraklion Crete [16]. Of 1363 singleton live births in the Rhea study, we included 905 pregnant women who provided information on first trimester diet via a validated semi-quantitative food frequency, after excluding 11 women with implausible values of energy intake (<600 or >6000 calories/day). Of the remaining 895 mother-child pairs, 570 participated at the 4 year follow up visit (median 4.2 years, IQR: 4.1–4.3), during which we measured anthropometry and collected non-fasting blood samples from 569 children, who comprised the population included in this analysis.
In both study populations, all procedures were in accordance with the ethical standards for human experimentation established by the Declaration of Helsinki and all women provided written informed consent. Institutional review boards of participating institutions approved each study.
Dietary intake during pregnancy
In Project Viva, mothers reported their diet since the time of their last menstrual period at study enrollment (median 9.9 weeks gestation), using a validated semi-quantitative food frequency questionnaire (FFQ) [17]. Rhea participants completed a validated FFQ at mean 14.6 weeks gestation [18]. To evaluate adherence to the Mediterranean diet during pregnancy, we used a score modified from one applied in a large cohort study in adults (Mediterranean Diet Score, MDS) [5]. To use the same thresholds for both cohorts, we applied cut-offs based where possible on current recommendations for pregnant women (Table S1) [19]. For components presumed to be beneficial (vegetables, fruits, fish, dairy products, legumes, whole grain products, nuts, and monounsatured fatty acids), women whose consumption was above recommendations were assigned a value of 1, otherwise they were assigned 0 points for intake equal or below the threshold. For components presumed to be detrimental (red and processed meat), we assigned 1 point for women whose consumption was below or equal to the threshold, and 0 for women whose consumption was above the threshold. The MDS thus ranged from 0 (minimal adherence to the Mediterranean diet) to 9 (maximal adherence). We have presented our results per 3 points increment on the MDS as conventionally a score of 0–3 represents low adherence, a score of 4–6 represents moderate adherence, and a score of 7–9 represents high adherence to the Mediterranean diet [5].
Child adiposity measures
In both cohorts, trained research assistants measured children’s weight, height, waist circumference, and subscapular (SS) and triceps (TR) skinfold thicknesses (Supplementary Methods).
Child blood pressure and cardiometabolic biomarkers
In both cohorts, trained research assistants measured systolic and diastolic blood pressure using a Dinamap automated oscillometric recorder (Supplementary Methods).
We collected blood via venipuncture and measured lipids [total cholesterol, and high-density lipoprotein cholesterol (HDL)], plasma leptin and adiponectin concentrations, and C-reactive protein (CRP) concentrations following standard protocols (Supplementary Methods).
Statistical Analysis
We used linear regression to estimate associations of adherence to the Mediterranean diet during pregnancy with adiposity or cardiometabolic outcomes in childhood. Generalized additive models (GAMs) were applied to explore the shape of the relationships between MDS and outcomes under study.
To select the confounders for adjustment in multivariable models, we used a directed acyclic graph approach based on prior knowledge about parental and child covariates that may be related to child adiposity and/or adherence to the Mediterranean diet in pregnancy. According to this graph (Supporting information Figure S1) we included the following variables in multivariable models: The first model was adjusted for the child’s sex and age at outcome measurement (crude model); the second model (confounder model) was additionally adjusted for maternal age at recruitment (years), education (high level: university or technical college degree), ethnicity (Greek/non Greek; USA/non USA citizen), race (black, Asian, Hispanic, white, other), pre-pregnancy body mass index [based on measured height at recruitment and prepregnancy self-reported weight (BMI, kg/m2)], smoking in pregnancy (never, quit before pregnancy, smoked during pregnancy), and parity (nulliparous; multiparous). In a third model (mediation model), we additionally adjusted for birth weight for gestational length z score and breastfeeding duration (months); and in a fourth model (also mediation) we additionally adjusted for child lifestyle characteristics [fast food intake (times/week, questionnaire based), TV viewing (hours/day, questionnaire based)] and child’s anthropometry at age of outcome assessment [height (for waist circumference and SS+TR) and child’s BMI (for cardiometabolic outcomes)] as potential mediators. We considered the confounder model as the main model.
We used a two-stage approach to assess the association of adherence to the Mediterranean diet during pregnancy with adiposity and cardiometabolic traits in children. First, we analyzed associations separately for each cohort. Second, we calculated pooled effect estimates using mixed models, including cohort and child age at outcome assessment as random effects and all other covariates as fixed effects. Finally, the overall summary effect of the individual cohorts was estimated using random effects meta-analysis to check the consistency with the pooled analysis. We presented results as combined estimates from the random-effect models with their 95% CIs.
We assessed effect modification by maternal pre-pregnancy BMI (≥ 25 versus <25 kg/m2), maternal smoking during pregnancy (yes versus no), and breastfeeding duration (>3 versus ≤ 3 months) through inclusion of the interaction terms in the models (statistically significant effect modification if p-value<0.05). We also performed further adjustment for energy intake in pregnancy (kcals/day), gestational diabetes (questionnaire based), and gestational weight gain (questionnaire based) for mother-child pairs with available information on these variables. We performed analyses with SAS version 9.3 software (SAS Institute, Cary, NC) and R version R3.1.
RESULTS
Participant characteristics and compliance with Mediterranean Diet
Intake of all food groups in the MDS during pregnancy differed between the two cohorts except nut intake (Supporting Information Figure S2). Pregnant women in Rhea cohort had higher intakes of almost all protective components of the MDS except legumes. On the other hand, they also reported a higher intake of red and processed meat products than Project Viva mothers. The MDS was higher in the Rhea cohort (Mean 3.8, SD 1.7) than in Project Viva (Mean 2.7, SD 1.6, p<0.001).
At the time of outcome assessment, Viva children were approximately 3.5 years older than those in Rhea; accordingly BMI and fat measurements were higher for the Viva children (Mean BMI, (SD): Viva; 17.1, (2.9); Rhea; 16.4 (1.9), p<0.01). Table S2 in the Supporting Information shows that in both cohorts, mothers without offspring follow-up data were more likely to be younger, smokers, less educated, and of non-white or non-Greek race/ethnicity accordingly.
Mediterranaean diet adherence and offspring obesity
GAMs examining the shape of the relationships of MDS with child z-BMI showed no departures from linearity overall and separately in Project Viva and Rhea cohorts (Supporting Information Figure S3).
Associations of MDS in pregnancy with offspring adiposity were broadly similar in each of the two cohorts studied separately (Table 2), although the magnitude and precision of estimates differed slightly. For example, in the covariate-adjusted model, each 3-point increment in MDS was associated with 0.13 units lower BMI z-score (95% CI: −0.24, −0.02) in Viva and 0.15 units lower in Rhea (95% CI: −0.29, 0.00). In the pooled analysis, for each 3-point increment in the MDS in pregnancy, offspring BMI z score was lower by 0.14 units (95% CI, −0.15 to −0.13), waist circumference was lower by 0.39 cm (95% CI, −0.64 to −0.14), and the sum of subscapular and triceps skin fold thicknesses was lower by 0.63mm (95% CI, −0.98 to −0.28) (Table 2).
TABLE 2.
Associations of Mediterranean diet score in pregnancy (per 3-point increment on a 0–9 scale) with offspring adiposity outcomes in Project Viva (N =997) and Rhea (N =569) cohorts, separately and in pooled analysis
| Model 1 | Model 2 | Model 3 | Model 4 | |||
|---|---|---|---|---|---|---|
| Basic Modela | Confounder Modelb | Mediator Model-Birth & Infant characteristicsc | Mediator Model-Child life-styled | |||
|
|
||||||
| Mean (SD) | β (95% CI) | |||||
| Project Viva (N =997) | ||||||
| BMI z-score | 0.36 (1.0) | −0.24 (−0.35,−0.12) | −0.13 (−0.24,−0.02) | −0.13 (−0.24,−0.01) | −0.13 (−0.24,−0.01) | |
| Waist circumference (cm) | 59.7 (7.9) | −0.95 (−1.82,−0.07) | −0.22 (−1.08, 0.64) | −0.23 (−1.08, 0.63) | −0.42 (−1.21, 0.38) | |
| SS+TR (mm) | 19.5 (9.2) | −1.40 (−2.43,−0.36) | −0.39 (−1.41, 0.63) | −0.22 (−1.24, 0.81) | −0.41 (−1.42, 0.61) | |
| Rhea (N =569) | ||||||
| BMI z-score | −0.16 (1.0) | −0.13 (−0.27, 0.02) | −0.15 (−0.29, 0.00) | −0.12 (−0.27, 0.03) | −0.13 (−0.28, 0.02) | |
| Waist circumference (cm) | 53.5 (4.9) | −0.44 (−1.17, 0.29) | −0.58 (−1.32, 0.17) | −0.45 (−1.23, 0.33) | −0.34 (−1.00, 0.33) | |
| SS+TR (mm) | 17.0 (5.0) | −0.69 (−1.43, 0.05) | −0.86 (−1.62,−0.10) | −0.80 (−1.59,−0.01) | −0.88 (−1.63,−0.12) | |
| Combined effect-Pooled analysis | ||||||
| BMI z-score | −0.19 (−0.27,−0.12) | −0.14 (−0.15,−0.13) | −0.13 (−0.13,−0.12) | −0.12 (−0.13,−0.12) | ||
| Waist circumference (cm) | −0.73 (−1.08,−0.38) | −0.39 (−0.64,−0.14) | −0.34 (−0.49,−0.18) | −0.40 (−0.46,−0.34) | ||
| SS+TR (mm) | −1.13 (−1.59,−0.67) | −0.63 (−0.98,−0.28) | −0.53 (−0.97,−0.09) | −0.61 (−0.91,−0.31) | ||
Basic model includes child sex and age at outcome
Confounder model is basic model additionally adjusted for maternal age, pre-pregnancy body mass index, race/ethnicity, education level, parity, and smoking during pregnancy
Mediator model - Birth & Infant characteristics is confounder model additionally adjusted for birth weight for gestation age z-score and breastfeeding duration
Mediator model - Child life-style characteristics is confounder model additionally adjusted for fast food intake, TV viewing, and child’s height (only for waist circumference and SS+TR) at age of outcome assessment
SS+TR, sum of subscapular and triceps skin fold thickness
Mediterranean Diet adherence and offspring cardiometabolic traits
In the pooled analysis, for each 3-point increment in the MDS in pregnancy, offspring systolic blood pressure was lower by 1.03 mmHg (95% CI, −1.65 to −0.42), and diastolic blood pressure was lower by 0.57 mmHg (95% CI, −0.98 to −0.16) (Table 3). Higher adherence to Mediterranean diet was also associated with lower offspring serum (log) leptin levels (% change −6.0; 95% CI −8.5 to −3.5).
TABLE 3.
Associations of Mediterranean diet score in pregnancy (per 3-point increment on a 0–9 scale) with offspring cardiometabolic risk factors in Project Viva (N =997) and Rhea (N = 569) cohorts, separately, and in pooled analysis
| Model 1 | Model 2 | Model 3 | Model 4 | |||
|---|---|---|---|---|---|---|
| Basic Modela | Confounder Modelb | Mediator Model- Birth & Infan characteristicsc | Mediator Model - Child life-styled | |||
|
|
||||||
| Mean (SD) | β (95% CI) | |||||
| Project Viva (N =997) | ||||||
|
| ||||||
| SBP (mmHg) | 94.5 (8.7) | −1.58 (−2.59,−0.57) | −1.38 (−2.42,−0.35) | −1.53 (−2.60,−0.46) | −1.17 (−2.18,−0.17) | |
| DBP (mmHg) | 54.4 (5.7) | −0.83 (−1.50,−0.16) | −0.81 (−1.50,−0.13) | −0.91 (−1.62,−0.20) | −0.77 (−1.48,−0.05) | |
|
|
||||||
| Mean (SD) | % Change (95% CI) | |||||
|
|
||||||
| CRP (ml/dl) | 0.8 (2.1) | −8.0 (−28.9, 19.0) | −2.9 (−21.2, 34.5) | −1.2 (−23.1, 33.2) | −1.0 (−23.5, 28.1) | |
| HDL-Cholesterol (mg/dl) | 57.1 (13.7) | 0.9 (−2.9, 4.8) | 0.5 (−3.4, 4.5) | 0.3 (−3.7, 4.5) | 0.5 (−3.5, 4.6) | |
| Total Cholesterol (mg/dl) | 159.8 (27.0) | 2.8 (0.0, 5.6) | 2.3 (−0.6, 5.2) | 2.0 (−1.0, 5.0) | 2.4 (−0.6, 5.5) | |
| Adiponectin (ul/ml) | 15.6 (8.8) | 5.3 (−3.5, 14.9) | 1.7 (−7.1, 11.4) | 3.9 (−5.3, 13.9) | 3.5 (−5.7, 13.7) | |
| Leptin (ng/ml) | 5.8 (6.8) | −7.5 (−18.7, 5.3) | −4.2 (−16.3, 9.7) | −2.3 (−14.8, 12.0) | −2.9 (−14.2, 10.0) | |
|
| ||||||
| Rhea (N = 569) | Mean (SD) | β (95% CI) | ||||
|
| ||||||
| SBP (mmHg) | 90.7 (7.9) | −0.35 (−1.70, 1.00) | −0.50 (−1.88, 0.88) | −0.85 (−2.25, 0.55) | −0.38 (−1.71, 0.95) | |
| DBPa(mmHg) | 53.9 (5.3) | −0.12 (−1.02, 0.79) | −0.26 (−1.20, 0.68) | −0.39 (−1.35, 0.57) | −0.14 (−1.12, 0.83) | |
|
|
||||||
| Mean (SD) | % Change (95% CI) | |||||
|
|
||||||
| CRP (ml/dl) | 0.2 (0.7) | −13.4 (−31.1, 8.8) | −7.6 (−27.0, 17.0) | −10.1 (−29.5, 14.6) | −8.6 (−28.6, 16.9) | |
| HDL -Cholesterol (mg/dl) | 46.6 (10.5) | 1.3 (−2.4, 5.1) | 0.6 (−3.2, 4.6) | 0.4 (−3.6, 4.5) | 0.5 (−3.4, 4.7) | |
| Total Cholesterol (mg/dl) | 156.7 (28.7) | 0.0 (−2.9, 2.9) | −1.1 (−4.1, 2.0) | −1.0 (−4.0, 2.1) | −1.1 (−4.1, 2.1) | |
| Adiponectin (ul/ml) | 15.0 (8.6) | −6.6 (−15.2, 2.8) | −5.7 (−14.8, 4.5) | −6.8 (−16.1, 3.6) | −3.7 (−13.3, 7.0) | |
| Leptin (ng/ml) | 2.8 (3.7) | −5.7 (−17.4, 7.8) | −7.7 (−19.5, 5.9) | −9.4 (−21.4, 4.6) | −4.0 (−14.1, 7.3) | |
|
| ||||||
| Combined effect-Pooled analysis | β (95% CI) | |||||
|
| ||||||
| SBP (mmHg) | −1.13 (−1.93,−0.33) | −1.03 (−1.65,−0.42) | −1.22 (−1.72,−0.71) | −0.82 (−1.40,−0.25) | ||
| DBP (mmHg) | −0.56 (−1.04,−0.07) | −0.57 (−0.98,−0.16) | −0.66 (−1.06,−0.27) | −0.49 (−0.96,−0.02) | ||
|
|
||||||
| % Change (95% CI) | ||||||
|
|
||||||
| CRP (ml/dl) | −11.9 (−15.8, −7.8) | −4.0 (−11.2, 3.8) | −6.4 (−13.9, 1.9) | −4.7 (−9.5, 0.3) | ||
| HDL -Cholesterol (mg/dl) | 1.1 (0.8, 1.4) | 0.6 (0.4, 0.8) | 0.6 (0.3, 0.8) | 0.2 (0.1, 0.2) | ||
| Total Cholesterol (mg/dl) | 1.4 (−0.5, 3.4) | 0.9 (−1.5, 3.3) | 0.9 (−1.4, 3.2) | 1.0 (−1.7, 3.7) | ||
| Adiponectin (ul/ml) | −0.9 (−8.7, 7.5) | −1.6 (−6.5, 3.4) | −0.9 (−8.0, 6.7) | −0.1 (−5.2, 5.2) | ||
| Leptin (ng/ml) | −6.4 (−7.5, −5.3) | −6.0 (−8.5, −3.5) | −5.9 (−11.0, −0.5) | −3.7 (−4.3, −3.1) | ||
SBP, Systolic Blood Pressure; DBP, Diastolic Blood Pressure; HDL, high-density lipoprotein; CRP: C reactive protein
Lipids, leptin, adiponectin and CRP were log transformed to normalize their distributions. We calculated percent change by exponentiating beta coefficients, subtracting by 1 and multiplying by 100
Basic model includes child sex and age at outcome
Confounder model also includes maternal age, pre-pregnancy body mass index, race/ethnicity, education level, parity, smoking during pregnancy, and child sex and age at outcome assessment
Mediator model - Birth & Infant characteristics is confounder model additionally adjusted for birth weight for gestation age z-score and breastfeeding duration
Mediator model - Child life-style characteristics is confounder model additionally adjusted for fast food intake, TV viewing, and child’s BMI at age of outcome assessment
Stratified and Sensitivity analyses
The study-level meta-analysis showed similar effect estimates, though with wider confidence intervals as expected (Supporting Information Table S3, Table S4, Figure S4). We further performed additional analyses to explore the extent to which observed associations differed according to pre-specified maternal and child characteristics. We saw no evidence for a multiplicative interaction of adherence to the Mediterranean diet during pregnancy with pre-pregnancy BMI, maternal smoking during pregnancy, and breastfeeding (p for interaction 0.10 to 0.99). Further adjustment for birth characteristics and infant feeding did not substantively alter any of adjusted models for childhood outcomes (Tables 2 and 3, Model 3). Results were also similar when we additionally adjusted for child TV watching, fast food intake, and child anthropometry at age of the outcome assessment (Tables 2 and 3, Model 4) though with wider confidence intervals (data not shown). Adjustment for gestational weight gain, energy intake, and gestational diabetes did not modify the direction of associations, though effect estimates were slightly attenuated (data not shown). A sensitivity analysis using the original MDS (based on cohort-specific median intake of 9 food groups) gave similar results with the presented analysis (data not shown).
DISCUSSION
In this analysis of prospective data from two cohorts with different levels and predictors of adherence to the Mediterranean diet, women with a greater adherence to the Mediterranean diet in early pregnancy had offspring who were less adipose and had lower blood pressure in childhood. The results were broadly similar in each of the two cohorts studied separately after controlling for a variety of confounders, remained robust in the pooled analysis, and were not mediated by several birth, infant and child characteristics.
Few human studies have evaluated the role of maternal diet for childhood obesity risk. A pregnancy cohort study from Spain showed that higher adherence to the Mediterranean diet in pregnancy was associated with lower waist circumference but not with BMI in preschool children [14], suggesting a specific effect on programming body fat distribution leading to a lower abdominal obesity risk without influencing general obesity. Two other pregnancy cohorts investigated the association between maternal dietary patterns and child body composition, and showed no significant associations [20,21]. Previous studies have focused on specific food groups and showed that the consumption of a high-meat, low-carbohydrate diet [22], as well as high fish intake during pregnancy [23], were associated with greater offspring adiposity.
In contrast with food groups or macronutrient analysis, the study of dietary patterns accounts for cumulative and interactive effects among nutrients, reflect real-world-dietary preferences, and may be particularly suitable for analysis in epidemiology of childhood obesity where many dietary components could be related with the outcome of interest [24]. A large cross-sectional study of 16,220 children aged 2–9 years in eight different European countries showed that a high adherence to the Mediterranean diet was inversely associated with overweight, obesity, and fat mass [9]. Suggested mechanisms include the low glycemic effect of Mediterranean diet and its high antioxidant content, which may lead to better glucose metabolism and metabolic function and finally influence individual susceptibly to weight gain later in life [4].
We also found that a greater adherence to the Mediterranean diet during pregnancy was associated with lower levels of systolic and diastolic blood pressure in childhood. A recent meta-analyses of clinical trials and prospective cohort studies in adults showed that a high adherence to the Mediterranean diet was associated with lower systolic and diastolic blood pressure [7]. Further, two observational cohort studies have showed that consumption of a high-meat, low-carbohydrate diet in pregnancy was associated higher adult blood pressure in the offspring [25,26]. These results were supported by animal studies suggesting that maternal high-fat diet can program rat offspring hypertension by activating the adipose renin-angiotensin system [27].
The pooled analysis revealed that higher adherence to the Mediterranean diet in early pregnancy was associated with lower leptin levels in childhood, likely reflecting lower offspring fat mass and lower triglyceride levels. A previous Project Viva analysis did not show association between Mediterranean diet adherence in pregnancy and leptin in cord blood [28], but we have previously reported that leptin in cord blood appears inversely associated with adiposity, whereas leptin in childhood predicts later excess weight gain.
Strengths of our study include the population-based prospective design, the fairly large sample size, the harmonized exposure estimates in the two cohorts, the detailed childhood body fat and cardiometabolic measurements, and the centralized statistical analysis following a consensus protocol. While in some respects the Rhea and Viva cohorts were ideally suited for pooling because of their similar data collection tools, there was also variability in confounders (such as maternal ethnicity and education level), age of outcome assessment, and interpretation of “serving” size between the two cohorts. In spite of all these sources of variability, we found fairly homogeneous effects across the different cohorts, indicating that our results are robust. We included women living in a Mediterranean country and in the USA therefore, our results can be generalized to other than Mediterranean settings.
This study also had some limitations. First, imperfections in dietary assessment are always a concern in nutritional epidemiology. However, in both cohorts, FFQs were validated, were only asking about a relatively short time period, which results in less recalling, estimation and abstraction for the participants, and the frequency scales used were almost identical in the two cohorts. When we further adjusted the final models for energy intake, results remained in the same direction, though attenuated. Indices like the MDS have inherent limitations such as assuming equal contribution from each component and variability in choosing cut-off points for each component. To minimize this variability, we used the same absolute cut-off point for each MDS component in both cohorts, so as to allow direct comparisons between the two cohorts. The levels of attrition in the Project Viva and Rhea cohorts are similar to those found in other birth cohort studies. We do not know obesity status of children lost to follow up. However, assuming that lost to follow up mother-child pairs, characterized by low socioeconomic status, may have a worse quality of diet during pregnancy and higher BMI in childhood, our estimates may be underestimated. We assessed several adiposity outcomes and cardiometabolic risk factors, raising concern about multiple testing. However, an application of Bonferroni correction to take into account multiple comparisons will be inappropriate in this case given that we are studying outcomes that are highly correlated [29]. We observed small effects for adiposity outcomes whereas the results for leptin and blood pressure were more powerful. Although a small decrement of 1-cm in child waist circumference might not be seem substantial at the individual level, the aggregative effect at the population level, as measured by a leftward shift in the distribution of abdominal obesity, may translate into a substantively large increase in the number of healthy children. Finally, it is important to note that this was an observational study and thus lacks the ability to determine causality.
CONCLUSION
In conclusion, our results from two pregnancy cohort studies in the USA and Greece, support the hypotheses that maternal adherence to the Mediterranean diet during pregnancy was associated with lower child adiposity, leptin and blood pressure levels. While intervention trials are needed to confirm these associations, it seems reasonable for health care providers to recommend healthy dietary patterns such as the Mediterranean Diet for pregnant women.
Supplementary Material
TABLE 1.
Maternal and child characteristics by Mediterranean diet score in Project Viva and Rhea cohorts
| Project Viva | Rhea | |||
|---|---|---|---|---|
|
|
||||
| MDS | MDS | |||
| No. (%) | Mean (SD) | No. (%) | Mean (SD) | |
| All | 997 (100) | 2.7 (1.6) | 569 (100) | 3.8 (1.7) |
| Maternal characteristics | ||||
| Age (years) | ||||
| <25 | 25 (2.5) | 1.7 (1.3) | 13 (2.3) | 4.4 (1.8) |
| 25–35 | 655 (65.7) | 2.7 (1.6) | 461 (81.4) | 3.7 (1.7) |
| ≥35 | 317 (31.8) | 2.9 (1.6) | 92 (16.3) | 3.9 (1.6) |
| Pre-pregnancy BMI ≥25 kg/m2 | ||||
| No | 651 (65.5) | 2.8 (1.6) | 377 (67.0) | 3.8 (1.6) |
| Yes | 343 (34.5) | 2.5 (1.5) | 186 (33.0) | 3.5 (1.8) |
| College graduate | ||||
| No | 289 (29.0) | 2.3 (1.4) | 381 (67.4) | 3.7 (1.7) |
| Yes | 708 (71.0) | 2.9 (1.6) | 184 (32.6) | 3.8 (1.7) |
| Race/ethnicity | NA | |||
| Black | 126 (12.6) | 2.3 (1.6) | ||
| Hispanic | 61 (6.1) | 2.8 (1.8) | ||
| Asian | 53 (5.3) | 2.8 (1.6) | ||
| White | 713 (71.5) | 2.8 (1.6) | 569 (100) | 3.8 (1.7) |
| Other | 44 (4.4) | 2.4 (1.7) | ||
| Married or cohabitating | ||||
| No | 71 (7.1) | 2.1 (1.6) | 11 (1.9) | 3.8 (1.9) |
| Yes | 925 (92.9) | 2.8 (1.6) | 558 (98.1) | 3.7 (1.7) |
| Nulliparous | ||||
| No | 516 (51.8) | 2.6 (1.6) | 316 (57.8) | 3.6 (1.7) |
| Yes | 481 (48.2) | 2.8 (1.6) | 231 (42.2) | 3.9 (1.7) |
| Smoking status | ||||
| Never | 693 (69.6) | 2.8 (1.6) | 364 (65.8) | 3.8 (1.7) |
| Quit before pregnancy | 209 (21.0) | 2.7 (1.5) | 97 (17.5) | 3.7 (1.6) |
| Smoked during pregnancy | 93 (9.3) | 2.2 (1.4) | 92 (16.6) | 3.6 (1.7) |
| Child characteristics | ||||
| Sex | ||||
| Male | 491 (49.2) | 2.7 (1.6) | 308 (54.1) | 3.6 (1.7) |
| Female | 506 (50.8) | 2.7 (1.6) | 261 (45.9) | 3.9 (1.7) |
| Gestation length | ||||
| <34 weeks | 15 (1.5) | 2.7 (1.4) | 8 (1.4) | 3.6 (1.9) |
| ≥34 weeks | 982 (98.5) | 2.7 (1.6) | 553 (98.6) | 3.8 (1.7) |
| Breastfeeding duration | ||||
| <3 months | 243 (25.7) | 2.4 (1.5) | 272 (50.0) | 3.7 (1.8) |
| ≥3 months | 701 (74.3) | 2.9 (1.6) | 272 (50.0) | 3.8 (1.6) |
MDS, Mediterranean diet score
Acknowledgments
We are grateful to all the participating families in USA and Greece who take part in these ongoing cohort studies.
Sources of financial support: The Rhea project was financially supported by European projects (EU FP6-2003-Food-3-NewGeneris, EU FP6. STREP Hiwate, EU FP7 ENV.2007.1.2.2.2. Project No 211250 Escape, EU FP7-2008-ENV-1.2.1.4 Envirogenomarkers, EU FP7-HEALTH-2009-single stage CHICOS, EU FP7 ENV.2008.1.2.1.6. Proposal No 226285 ENRIECO, EU-FP7-HEALTH-2012 Proposal No 308333 HELIX, and the Greek Ministry of Health (Program of Prevention of obesity and neurodevelopmental disorders in preschool children, in Heraklion district, Crete, Greece: 2011–2014; “Rhea Plus”: Primary Prevention Program of Environmental Risk Factors for Reproductive Health, and Child Health: 2012-15). Dr Chatzi was additionally supported by Fulbright institution. Project Viva was supported by grants from the US National Institutes of Health (R01 HL075504, R37 HD 034568, R01 ES016314). Dr. Oken was additionally funded by K24 HD069408, P30 DK040561, and P30 DK092924.
Footnotes
CONTRIBUTORS
Designed research, LC, EO
Collected data, provided essential materials (cohort specific databases necessary for research), SR, SK, GC, AM, KS, MV.
Analyzed data and performed statistical analysis, SR, VG.
Wrote paper, LC, KEJ, EO.
Had primary responsibility for final content, all authors.
Study supervision: LC, MK, CM, MG, EO.
All authors read and approved the final manuscript.
CONFLICT OF INTEREST STATEMENT
None of the authors had a conflict of interest to report.
References
- 1.Gluckman PD, Hanson MA, Cooper C, Thornburg KL. Effect of in utero and early-life conditions on adult health and disease. N Engl J Med. 2008;359(1):61–73. doi: 10.1056/NEJMra0708473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Symonds ME, Sebert SP, Hyatt MA, Budge H. Nutritional programming of the metabolic syndrome. Nat Rev Endocrinol. 2009;5(11):604–10. doi: 10.1038/nrendo.2009.195. [DOI] [PubMed] [Google Scholar]
- 3.Murrin C, Shrivastava A, Kelleher CC. Maternal macronutrient intake during pregnancy and 5 years postpartum and associations with child weight status aged five. Eur J Clin Nutr. 2013;67(6):670–9. doi: 10.1038/ejcn.2013.76. [DOI] [PubMed] [Google Scholar]
- 4.Donnelly JM, Walsh JM, Byrne J, Molloy EJ, McAuliffe FM. Impact of maternal diet on neonatal anthropometry: a randomized controlled trial. Pediatr Obes. 2015;10(1):52–6. doi: 10.1111/j.2047-6310.2013.00216.x. [DOI] [PubMed] [Google Scholar]
- 5.Trichopoulou A, Costacou T, Bamia C, Trichopoulos D. Adherence to a Mediterranean diet and survival in a Greek population. N Engl J Med. 2003;348(26):2599–608. doi: 10.1056/NEJMoa025039. [DOI] [PubMed] [Google Scholar]
- 6.Huo R, Du T, Xu Y, Xu W, Chen X, Sun K, Yu X. Effects of Mediterranean-style diet on glycemic control, weight loss and cardiovascular risk factors among type 2 diabetes individuals: a meta-analysis. Eur J Clin Nutr. 2015;69(11):1200–8. doi: 10.1038/ejcn.2014.243. [DOI] [PubMed] [Google Scholar]
- 7.Kastorini CM, Milionis HJ, Esposito K, Giugliano D, Goudevenos JA, Panagiotakos DB. The effect of Mediterranean diet on metabolic syndrome and its components: a meta-analysis of 50 studies and 534,906 individuals. J Am Coll Cardiol. 2011;57(11):1299–313. doi: 10.1016/j.jacc.2010.09.073. [DOI] [PubMed] [Google Scholar]
- 8.Romaguera D, Norat T, Vergnaud AC, Mouw T, May AM, Agudo A, Buckland G, Slimani N, Rinaldi S, Couto E, Clavel-Chapelon F, Boutron-Ruault MC, Cottet V, Rohrmann S, Teucher B, Bergmann M, Boeing H, Tjonneland A, Halkjaer J, Jakobsen MU, Dahm CC, Travier N, Rodriguez L, Sanchez MJ, Amiano P, Barricarte A, Huerta JM, Luan J, Wareham N, Key TJ, Spencer EA, Orfanos P, Naska A, Trichopoulou A, Palli D, Agnoli C, Mattiello A, Tumino R, Vineis P, Bueno-de-Mesquita HB, Buchner FL, Manjer J, Wirfalt E, Johansson I, Hellstrom V, Lund E, Braaten T, Engeset D, Odysseos A, Riboli E, Peeters PH. Mediterranean dietary patterns and prospective weight change in participants of the EPIC-PANACEA project. Am J Clin Nutr. 2010;92(4):912–21. doi: 10.3945/ajcn.2010.29482. [DOI] [PubMed] [Google Scholar]
- 9.Tognon G, Hebestreit A, Lanfer A, Moreno LA, Pala V, Siani A, Tornaritis M, De Henauw S, Veidebaum T, Molnar D, Ahrens W, Lissner L. Mediterranean diet, overweight and body composition in children from eight European countries: cross-sectional and prospective results from the IDEFICS study. Nutr Metab Cardiovasc Dis. 2014;24(2):205–13. doi: 10.1016/j.numecd.2013.04.013. [DOI] [PubMed] [Google Scholar]
- 10.Saunders L, Guldner L, Costet N, Kadhel P, Rouget F, Monfort C, Thome JP, Multigner L, Cordier S. Effect of a Mediterranean diet during pregnancy on fetal growth and preterm delivery: results from a French Caribbean Mother-Child Cohort Study (TIMOUN) Paediatr Perinat Epidemiol. 2014;28(3):235–44. doi: 10.1111/ppe.12113. [DOI] [PubMed] [Google Scholar]
- 11.Mikkelsen TB, Osterdal ML, Knudsen VK, Haugen M, Meltzer HM, Bakketeig L, Olsen SF. Association between a Mediterranean-type diet and risk of preterm birth among Danish women: a prospective cohort study. Acta Obstet Gynecol Scand. 2008;87(3):325–30. doi: 10.1080/00016340801899347. [DOI] [PubMed] [Google Scholar]
- 12.Timmermans S, Steegers-Theunissen RP, Vujkovic M, den Breeijen H, Russcher H, Lindemans J, Mackenbach J, Hofman A, Lesaffre EE, Jaddoe VV, Steegers EA. The Mediterranean diet and fetal size parameters: the Generation R Study. Br J Nutr. 2012;108(8):1399–409. doi: 10.1017/S000711451100691X. [DOI] [PubMed] [Google Scholar]
- 13.Chatzi L, Mendez M, Garcia R, Roumeliotaki T, Ibarluzea J, Tardon A, Amiano P, Lertxundi A, Iniguez C, Vioque J, Kogevinas M, Sunyer J. Mediterranean diet adherence during pregnancy and fetal growth: INMA (Spain) and RHEA (Greece) mother-child cohort studies. Br J Nutr. 2012;107(1):135–45. doi: 10.1017/S0007114511002625. [DOI] [PubMed] [Google Scholar]
- 14.Fernandez-Barres S, Romaguera D, Valvi D, Martinez D, Vioque J, Navarrete-Munoz EM, Amiano P, Gonzalez-Palacios S, Guxens M, Pereda E, Riano I, Tardon A, Iniguez C, Arija V, Sunyer J, Vrijheid M. Mediterranean dietary pattern in pregnant women and offspring risk of overweight and abdominal obesity in early childhood: the INMA birth cohort study. Pediatr Obes. 2016 doi: 10.1111/ijpo.12092. [DOI] [PubMed] [Google Scholar]
- 15.Oken E, Baccarelli AA, Gold DR, Kleinman KP, Litonjua AA, De Meo D, Rich-Edwards JW, Rifas-Shiman SL, Sagiv S, Taveras EM, Weiss ST, Belfort MB, Burris HH, Camargo CA, Jr, Huh SY, Mantzoros C, Parker MG, Gillman MW. Cohort profile: project viva. Int J Epidemiol. 2015;44(1):37–48. doi: 10.1093/ije/dyu008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Chatzi L, Plana E, Daraki V, Karakosta P, Alegkakis D, Tsatsanis C, Kafatos A, Koutis A, Kogevinas M. Metabolic syndrome in early pregnancy and risk of preterm birth. Am J Epidemiol. 2009;170(7):829–36. doi: 10.1093/aje/kwp211. [DOI] [PubMed] [Google Scholar]
- 17.Fawzi WW, Rifas-Shiman SL, Rich-Edwards JW, Willett WC, Gillman MW. Calibration of a semi-quantitative food frequency questionnaire in early pregnancy. Ann Epidemiol. 2004;14(10):754–62. doi: 10.1016/j.annepidem.2004.03.001. [DOI] [PubMed] [Google Scholar]
- 18.Chatzi L, Melaki V, Sarri K, Apostolaki I, Roumeliotaki T, Georgiou V, Vassilaki M, Koutis A, Bitsios P, Kogevinas M. Dietary patterns during pregnancy and the risk of postpartum depression: the mother-child ‘Rhea’ cohort in Crete, Greece. Public Health Nutr. 2011;14(9):1663–70. doi: 10.1017/S1368980010003629. [DOI] [PubMed] [Google Scholar]
- 19.McGuire S. U.S. Department of Agriculture and U.S. Department of Health and Human Services, Dietary Guidelines for Americans, 2010. 7th Edition, Washington, DC: U.S. Government Printing Office, January 2011. Adv Nutr. 2011;2(3):293–4. doi: 10.3945/an.111.000430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Poon AK, Yeung E, Boghossian N, Albert PS, Zhang C. Maternal dietary patterns during third trimester in association with birthweight characteristics and early infant growth. Scientifica (Cairo) 2013;2013:786409. doi: 10.1155/2013/786409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.van den Broek M, Leermakers ET, Jaddoe VW, Steegers EA, Rivadeneira F, Raat H, Hofman A, Franco OH, Kiefte-de Jong JC. Maternal dietary patterns during pregnancy and body composition of the child at age 6 y: the Generation R Study. Am J Clin Nutr. 2015;102(4):873–80. doi: 10.3945/ajcn.114.102905. [DOI] [PubMed] [Google Scholar]
- 22.Yin J, Quinn S, Dwyer T, Ponsonby AL, Jones G. Maternal diet, breastfeeding and adolescent body composition: a 16-year prospective study. Eur J Clin Nutr. 2012;66(12):1329–34. doi: 10.1038/ejcn.2012.122. [DOI] [PubMed] [Google Scholar]
- 23.Stratakis N, Roumeliotaki T, Oken E, Barros H, Basterrechea M, Charles MA, Eggesbo M, Forastiere F, Gaillard R, Gehring U, Govarts E, Hanke W, Heude B, Iszatt N, Jaddoe VW, Kelleher C, Mommers M, Murcia M, Oliveira A, Pizzi C, Polanska K, Porta D, Richiardi L, Rifas-Shiman SL, Schoeters G, Sunyer J, Thijs C, Viljoen K, Vrijheid M, Vrijkotte TG, Wijga AH, Zeegers MP, Kogevinas M, Chatzi L. Fish Intake in Pregnancy and Child Growth: A Pooled Analysis of 15 European and US Birth Cohorts. JAMA Pediatr. 2016;170(4):381–90. doi: 10.1001/jamapediatrics.2015.4430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hu FB. Dietary pattern analysis: a new direction in nutritional epidemiology. Curr Opin Lipidol. 2002;13(1):3–9. doi: 10.1097/00041433-200202000-00002. [DOI] [PubMed] [Google Scholar]
- 25.Campbell DM, Hall MH, Barker DJ, Cross J, Shiell AW, Godfrey KM. Diet in pregnancy and the offspring’s blood pressure 40 years later. Br J Obstet Gynaecol. 1996;103(3):273–80. doi: 10.1111/j.1471-0528.1996.tb09718.x. [DOI] [PubMed] [Google Scholar]
- 26.Shiell AW, Campbell-Brown M, Haselden S, Robinson S, Godfrey KM, Barker DJ. High-meat, low-carbohydrate diet in pregnancy: relation to adult blood pressure in the offspring. Hypertension. 2001;38(6):1282–8. doi: 10.1161/hy1101.095332. [DOI] [PubMed] [Google Scholar]
- 27.Guberman C, Jellyman JK, Han G, Ross MG, Desai M. Maternal high-fat diet programs rat offspring hypertension and activates the adipose renin-angiotensin system. Am J Obstet Gynecol. 2013;209(3):262 e1–8. doi: 10.1016/j.ajog.2013.05.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Mantzoros CS, Sweeney L, Williams CJ, Oken E, Kelesidis T, Rifas-Shiman SL, Gillman MW. Maternal diet and cord blood leptin and adiponectin concentrations at birth. Clin Nutr. 2010;29(5):622–6. doi: 10.1016/j.clnu.2010.03.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiology. 1990;1(1):43–6. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.