Maternal Dietary Patterns during Pregnancy Are Associated with Newborn Body Composition

Abstract

Background: Maternal dietary intake during pregnancy may influence offspring growth and adiposity. Specific dietary patterns associated with newborn adiposity have not been identified.

Objective: We aimed to identify patterns of maternal dietary intake associated with gestational weight gain (GWG) and fasting glucose during pregnancy and to evaluate whether adherence to these patterns is associated with newborn adiposity.

Methods: In the Healthy Start prospective cohort, dietary intake during pregnancy was assessed via 24-h recalls. Reduced-rank regression identified dietary patterns predictive of GWG and fasting glucose. Associations between dietary patterns and newborn fat mass, fat-free mass, and adiposity were estimated by using linear regression models among 764 ethnically diverse mother-infant pairs.

Results: Two dietary patterns were identified. Pattern 1, correlated with greater GWG (r = 0.22, P < 0.01), was characterized by a higher consumption of poultry, nuts, cheese, fruits, whole grains, added sugars, and solid fats. Greater adherence to pattern 1 (upper compared with lower tertile) predicted a greater newborn fat-free mass (61 g; 95% CI: 12, 110 g) but no difference in fat mass or adiposity. Pattern 2, correlated with greater maternal fasting glucose (r = 0.16, P < 0.01), was characterized by a higher consumption of eggs, starchy vegetables, solid fats, fruits, and nonwhole grains and a lower consumption of dairy foods, dark-green vegetables, and whole grains. Greater adherence to pattern 2 was associated with a greater newborn birth weight (80 g; 95% CI: 15, 145 g), fat mass (33 g; 95% CI: 8, 59 g), and adiposity (0.9%; 95% CI: 0.3%, 1.6%).

Conclusions: Among pregnant women, adherence to a dietary pattern characterized by an intake of poultry, nuts, cheese, and whole grains was associated with greater GWG but not maternal fasting glucose or newborn adiposity. Adherence to a pattern characterized by an intake of eggs, starchy vegetables, and nonwhole grains was associated with higher maternal fasting glucose and greater newborn adiposity. Maternal dietary patterns during pregnancy may influence newborn body composition.

Introduction

The nutritional and metabolic environment of pregnancy is believed to have lasting effects on the health and chronic disease risk of the offspring (1, 2). The status of maternal nutrition when entering pregnancy, as indicated by BMI (in kg/m²), is associated with newborn adiposity and the risk of overweight and obesity later in life (3–5). Maternal dietary intake during pregnancy may influence gestational weight gain (GWG) and metabolic changes leading to fetal overnutrition (6–11). Recently, nutritional epidemiologists have focused on dietary patterns, rather than specific macro- or micronutrients, to examine the effects of the diet on health. Dietary patterns have the benefit of examining foods consumed in the context of the overall diet, allowing for interactive effects between dietary components and accommodating the high correlations that may occur between specific foods or nutrients (12).

In pregnancy, both data-driven and a priori dietary patterns have been associated with birth outcomes, including preterm delivery (13–15) and birth weight (16, 17). We have previously reported that adherence to the USDA’s Healthy Eating Index was associated with lower newborn adiposity in the Healthy Start study (18). In some other studies, however, maternal dietary patterns were not independently associated with offspring size or body composition (19–21). Inconsistent associations between maternal diet and offspring adiposity may result from the fact that the specific dietary patterns most strongly associated with newborn body composition have not been described. The method of reduced-rank regression allows the creation of dietary patterns to maximize the variation explained in intermediate variables proposed to be along the causal pathway between dietary intake and the health outcome of interest (22). In this study, we used reduced-rank regression to identify dietary patterns associated with maternal GWG and midpregnancy fasting glucose, 2 factors previously identified as significant predictors of newborn adiposity in this population (3, 23). We hypothesized that dietary patterns associated with greater maternal GWG and glucose would predict greater newborn fat mass and adiposity.

Methods

Study design and participants.

Participants were enrolled in Healthy Start, a prospective cohort study that recruited 1410 pregnant women from outpatient prenatal clinics at the University of Colorado Hospital from 2009 to 2014. Eligible participants were ≥16 y old with a singleton pregnancy, had completed <24 wk of gestation at the time of enrollment, had no previous history of stillbirth or preterm birth at <25 wk of gestation, and had no preexisting diabetes, cancer, psychiatric illness, or asthma treated with steroid medications. Participants completed 2 in-person visits during pregnancy, at a median of 17 wk (range: 10–24 wk) and 27 wk (range: 24–32 wk) of gestation, and 1 visit at delivery. Study protocols were approved by the Colorado Multiple Institutional Review Board. Written, informed consent was obtained before the first study visit. Participants with missing data on exposures, outcomes, or covariates were excluded from this analysis.

Dietary assessment.

At the first study visit, participants received instruction from trained study staff on how to complete monthly web-based dietary recalls using the National Cancer Institute’s Automated Self-Administered 24-h Recall (ASA24) (24). The validity of the ASA24 has been reported as comparable to that of the USDA’s Automated Multi-Pass Method, capturing ∼80% of observed foods consumed (25). The ASA24 software calculated quantities consumed per day from each of the USDA’s MyPyramid Equivalent food groups (26) by using the dietary intake data. The Diet, Physical Activity and Body Composition Core of the Nutrition Obesity Research Center at the University of North Carolina at Chapel Hill assisted in the dietary data collection and processing.

Intermediate variables.

Maternal weight during pregnancy was measured at each in-person study visit, and all weights measured at clinic visits were abstracted from the prenatal medical record by study personnel. Maternal weight before pregnancy was obtained either from medical records (∼90%) or from participant self-report at the first study visit (∼10%). GWG was calculated by subtracting the prepregnancy weight from the last recorded maternal weight during pregnancy. The median gestational age at the last recorded weight during pregnancy was 271 d (IQR: 262–278 d), and the median number of days between the last recorded weight and delivery was 5 d (IQR 2–8 d). Glucose concentration was measured in a fasting serum sample collected at a median of 27 wk of gestation by the Core Laboratory of the University of Colorado Clinical and Translational Research Centers with the use of an AU480 Chemistry Analyzer with a hexokinase reaction (Beckman Coulter).

Body composition measurement.

Newborn body composition was measured within 3 d of birth by using air displacement plethysmography via the PEA POD device (COSMED). The PEA POD estimates fat mass and fat-free mass from directly measured body mass and volume by using whole-body densitometry and has been shown to produce estimates of adiposity that are reproducible and not significantly different from the reference 4-compartment method (27). Adiposity was calculated as fat mass divided by the sum of fat mass and fat-free mass. Body composition was measured ≥2 times for each infant by trained clinical personnel; a third measurement was conducted if the adiposity differed by >2%, and the average of the 2 closest measures was used.

Other variables.

Participants reported their age, race/ethnicity, education, and number of previous pregnancies via questionnaires at the first study visit. Additionally, smoking and physical activity during pregnancy were reported at each study visit for the preceding 3 mo of pregnancy. Physical activity was reported by using the validated Pregnancy Physical Activity Questionnaire (28). Infant birth weight, sex, and gestational age at birth were obtained from the medical record.

Statistical analysis.

Analysis was restricted to participants who completed ≥2 ASA24s during pregnancy in order to reduce measurement error in energy intake due to underreporting (29). Daily intakes from each of the MyPyramid Equivalent food groups were averaged over all recalls. Participants were excluded if their average energy intake was considered implausible (defined as <500 or >5000 kcal/d). Two food groups consumed by <25% of participants were excluded (organ meats, alcoholic beverages), and intakes of fish high in omega-3 FAs and fish low in omega-3 FAs were combined. Average intakes of the 24 remaining food groups were log transformed with zero values replaced by an arbitrarily small value of 0.001. The average intake was regressed on average daily intake of total energy (kilocalories), and residuals were used in subsequent analyses to examine the effects of dietary composition independent of total energy intake (30, 31).

Dietary patterns were obtained by entering the residual intakes from each food group into a reduced-rank regression model designed to simultaneously maximize the variability explained in 2 hypothesized intermediate variables: GWG and midpregnancy fasting glucose. These variables have been significantly associated with neonatal adiposity in previous studies in this cohort (3, 23) and therefore were considered plausible candidates for the pathway by which maternal dietary patterns may influence newborn adiposity. Two dietary patterns were obtained by this method. Factor scores for each of the dietary patterns were outputted for each individual and categorized into tertiles to examine nonlinear associations. A higher pattern score (and thus higher tertile) indicated a greater degree of adherence to that dietary pattern. Pearson’s correlation coefficients were estimated between each dietary pattern score and each intermediate. Trends across tertiles were examined by using Cochran-Armitage trend tests for binary variables and univariate linear regression models with tertile of dietary score as an ordinal predictor for continuous variables.

Separate multivariable linear regression models were used to estimate associations between tertiles of dietary pattern adherence and the following newborn outcomes: birth weight, fat mass, fat-free mass, and adiposity (fat mass as a percentage of total body mass). Covariates were selected a priori based on a directed acyclic graph indicating variables related to both maternal dietary intake and newborn adiposity but not on the causal pathway between the exposure and outcome. Covariates included in models for birth weight were: maternal age (years), maternal prepregnancy BMI, maternal race/ethnicity (non-Hispanic white, Hispanic, non-Hispanic black, and all others), maternal education completed (less than high school, high school or equivalent, some college or associate degree, 4-y college, and graduate degree), previous pregnancies (any compared with none), smoking during pregnancy (any compared with none), physical activity during pregnancy (quartiles), infant gestational age at birth (days), and infant sex. Models for body composition (fat mass, fat-free mass, and adiposity) were adjusted for the same covariates and additionally for the time elapsed between birth and the body composition measurement (days). Sensitivity analyses were performed to exclude participants with diagnosed gestational diabetes or those who delivered low-birth-weight infants (<2500 g). Additionally, patterns were redefined among participants with ≥3 ASA24s during pregnancy in order to examine the robustness of the findings. Analyses were conducted in SAS 9.4 (SAS Institute). Data are presented as means ± SDs.

Results

Of the 1410 women enrolled in the Healthy Start study, 11 withdrew from the study before delivery, and an additional 17 experienced fetal demise. Of the remaining 1382 participants, 1037 completed ≥2 ASA24s during pregnancy. Two participants were excluded because of implausible average energy intakes (<500 or >5000 kcal/d). Additionally, 4 participants were excluded because of missing GWG values, 56 because of missing glucose measurements, 186 because of missing data on offspring body composition in the first 3 d after birth, and 25 because of missing covariate data, resulting in an analytic sample of 764 mother-infant pairs.

Participating women were 60% non-Hispanic white, 21% Hispanic, 12% non-Hispanic black, and 6% of other racial and ethnic groups. The BMI among participants was 25.4 ± 5.9. The median number of ASA24s completed during pregnancy was 3 (range: 2–8), and the daily total energy intake was 2068 ± 608 kcal. Women who were included in the analysis were generally representative of the larger Healthy Start cohort, with the following exceptions: women in the analytic sample were more likely to be non-Hispanic white, more likely to have completed education beyond high school, and less likely to be <20 y old (Supplemental Table 1).

Two dietary patterns were obtained by using reduced-rank regression (Supplemental Table 2). These patterns explained 8.9% of the variability in the predictors (MyPyramid Equivalent food groups) and 5.1% of the variability in the intermediate variables (maternal fasting glucose and GWG; Supplemental Table 3). Food groups with high factor loadings (≥0.20) on dietary pattern 1 included certain healthy protein sources (poultry, nuts, and seeds), whole grains, cheese, citrus fruits, melons and berries, and other fruits, as well as added sugars and discretionary solid fat. Dietary pattern 2 was characterized by high factor loadings for eggs, potatoes, other starchy vegetables, discretionary solid fat, citrus, melons and berries, and non-whole grains. Dietary pattern 2 also had low factor loadings (≤−0.20) for yogurt, added sugars, soy products (tofu and meat analogs), dark-green vegetables, and whole grains. There were no food groups with low factor loadings (≤−0.20) for dietary pattern 1.

Characteristics of mothers and infants differed by the tertiles of each of the 2 dietary patterns. Women in the highest tertile of adherence to dietary pattern 1 had a lower average BMI and were more likely to be non-Hispanic white, to have more than a high school education, and to have no previous pregnancies compared with women in the lowest tertile of adherence (Table 1). By contrast, women in the highest tertile of adherence to dietary pattern 2 were younger on average, less likely to be non-Hispanic white, and less likely to have completed more than a high school education compared with women in the lowest tertile of adherence. With regard to the intermediate variables, dietary pattern 1 was correlated with greater GWG (r = 0.22, P < 0.01) but lower fasting glucose at midpregnancy (r = −0.14, P < 0.01). Dietary pattern 2 was correlated with greater fasting glucose (r = 0.16, P < 0.01) and less strongly correlated with greater GWG (r = 0.09, P = 0.02).

TABLE 1.

Maternal and infant characteristics among 764 participants in the Healthy Start study by tertile of maternal adherence to each dietary pattern¹

	Dietary pattern 1, tertile				Dietary pattern 2, tertile
Characteristic	1 (n = 239)	2 (n = 258)	3 (n = 267)	P-trend2	1 (n = 238)	2 (n = 267)	3 (n = 259)	P-trend2
Adherence score	−1.1 ± 0.7	0.1 ± 0.2	0.9 ± 0.3	<0.001	−0.9 ± 0.5	0.0 ± 0.2	0.9 ± 0.5	<0.001
Total energy intake, kcal/d	2088 ± 812	2066 ± 521	2051 ± 454	0.50	2044 ± 627	2089 ± 615	2068 ± 583	0.67
Maternal age, y	27.9 ± 6.4	29.4 ± 5.6	28.5 ± 5.5	0.34	29.5 ± 5.3	29.1 ± 5.8	27.4 ± 6.2	<0.001
Prepregnancy BMI, kg/m2	26.7 ± 6.6	24.7 ± 5.9	25.0 ± 5.1	0.002	25.0 ± 5.8	25.3 ± 6.1	25.9 ± 5.8	0.08
Non-Hispanic white	104 (44)	170 (66)	184 (69)	<0.001	176 (74)	160 (60)	122 (47)	<0.001
More than high school education	143 (60)	221 (86)	211 (79)	<0.001	190 (80)	212 (79)	173 (67)	<0.001
Any previous pregnancies	158 (66)	171 (66)	151 (57)	0.02	145 (61)	169 (63)	166 (64)	0.47
Any smoking during pregnancy	14 (6)	12 (5)	16 (6)	0.93	10 (4)	16 (6)	16 (6)	0.34
Physical activity during pregnancy, METs3/wk	186 ± 85	179 ± 71	183 ± 77	0.71	180 ± 75	185 ± 75	182 ± 83	0.72
Male infant	106 (44)	144 (56)	136 (51)	0.16	120 (50)	152 (57)	114 (44)	0.14
Gestational age at birth, d	276 ± 9	277 ± 8	278 ± 9	0.02	277 ± 9	277 ± 8	276 ± 8	0.54
Preterm	11 (5)	6 (2)	6 (2)	0.13	9 (4)	8 (3)	6 (2)	0.34
Gestational weight gain, kg	12.4 ± 6.6	14.7 ± 5.6	15.7 ± 5.7	<0.001	13.5 ± 6.1	14.7 ± 5.6	14.7 ± 6.5	0.03
Fasting glucose, mg/dL	79.9 ± 10.6	76.8 ± 6.3	76.8 ± 6.6	<0.001	76.1 ± 8.2	77.7 ± 7.8	79.4 ± 8.1	<0.001
Birth weight, g	3243 ± 440	3275 ± 429	3322 ± 423	0.04	3251 ± 457	3301 ± 425	3289 ± 413	0.34
Newborn fat mass, g	302 ± 144	283 ± 146	286 ± 156	0.26	275 ± 153	289 ± 148	304 ± 145	0.03
Newborn fat-free mass, g	2793 ± 336	2836 ± 335	2880 ± 324	0.003	2827 ± 352	2852 ± 319	2834 ± 329	0.82
Newborn adiposity, %	9.5 ± 3.9	8.8 ± 3.8	8.8 ± 4.0	0.04	8.6 ± 4.0	8.9 ± 3.9	9.5 ± 3.9	0.01

¹Values are means ± SDs or n (%).

²Obtained from Cochran-Armitage trend tests for binary variables or univariate linear regression for continuous variables.

³MET, metabolic equivalent.

In covariate-adjusted analyses, adherence to dietary pattern 1 was associated with greater newborn fat-free mass but not fat mass (Table 2). Women in the highest tertile of adherence to dietary pattern 1 had newborns with 61 g higher fat-free mass at birth (95% CI: 12, 110 g) compared with women in the lowest tertile. By contrast, adherence to dietary pattern 2 was associated with greater newborn fat mass, birth weight, and adiposity. Women in the highest tertile of adherence to dietary pattern 2 had offspring with 33 g higher fat mass (95% CI: 8, 59 g), 80 g higher birth weight (95% CI: 15, 145 g), and 0.9% higher adiposity at birth (95% CI: 0.3%, 1.6%) compared with women in the lowest tertile. Similar results were obtained in sensitivity analyses excluding 25 participants with diagnosed gestational diabetes or 23 participants with low birth weight (data not shown). The 2 dietary patterns identified were similar when analyses were restricted to 528 women with ≥3 ASA24s during pregnancy.

TABLE 2.

Maternal dietary pattern adherence and differences in offspring birth weight and body composition among 764 participants in the Healthy Start study, 2009–2014¹

Dietary pattern, tertile	Birth weight, g	Adiposity, %	Fat mass, g	Fat-free mass, g
Pattern 1
1 (n = 239)	Ref	Ref	Ref	Ref
2 (n = 258)	12.0 (−54.4, 78.3)	−0.39 (−1.09, 0.30)	−11.4 (−37.6, 14.7)	17.1 (−32.1, 66.4)
3 (n = 267)	52.6 (−13.5, 118.6)	−0.57 (−1.26, 0.12)	−12.4 (−38.4, 13.6)	60.6 (11.6, 109.6)
Pattern 2
1 (n = 238)	Ref	Ref	Ref	Ref
2 (n = 267)	54.8 (−8.1, 117.6)	0.56 (−0.09, 1.22)	21.1 (−3.6, 45.8)	23.2 (−23.6, 70.1)
3 (n = 259)	80.2 (15.4, 144.9)	0.94 (0.26, 1.61)	33.4 (8.0, 58.8)	38.0 (−10.3, 86.2)

¹Values are mean differences (95% CIs) obtained from linear regression models adjusted for maternal age, prepregnancy BMI, race/ethnicity, education, parity, smoking, average weekly physical activity during pregnancy, infant sex, gestational age at birth, and postnatal age at body composition measurement (for fat mass, fat-free mass, and adiposity only). Ref, reference category.

Discussion

In an ethnically diverse cohort of pregnant women in Colorado, we identified 2 maternal dietary patterns during pregnancy that were associated with newborn body composition. A dietary pattern characterized by poultry, nuts, cheese, fruits, and whole grains was associated with greater GWG but lower maternal glucose and no differences in newborn adiposity. By contrast, a pattern characterized by starchy vegetables, eggs, non-whole grains, and a low intake of dairy, dark-green vegetables, whole grains, and soy was associated with greater maternal fasting glucose and greater newborn birth weight and adiposity. Of note, these associations were independent of other known predictors of newborn adiposity, including maternal prepregnancy BMI (3), daily total energy intake (32), and physical activity (33).

The dietary patterns we identified using reduced-rank regression did not clearly resemble “healthy” or “unhealthy” diets as defined by scales such as the USDA Healthy Eating Index. This is to be expected, because the patterns were defined to maximize the explained variability in maternal GWG and fasting glucose. Healthy eating recommendations from the USDA (e.g., choosemyplate.gov) include higher intakes of fruits, vegetables, and grains (with at least half of the intake from whole grains), moderate intakes of protein (such as low-fat meat and poultry, beans and legumes, nuts and seeds, or fish) and low-fat dairy products, and a limited intake of oils (with a preference for MUFA and PUFA), refined grains, and added sugars.

Dietary pattern 1, although characterized by a higher intake of certain recommended protein sources, such as poultry, nuts and seeds, and complex carbohydrates in whole grains and fruits, also included added sugars and solid fat (as from meat, dairy, or hydrogenated oils). Dietary pattern 2 was characterized by a high intake of carbohydrates (potatoes, other starchy vegetables, and non-whole grains) and solid fat but also certain fruits (citrus and melons and berries). Dietary pattern 2 may be more clearly defined by those recommended food groups with the lowest factor loadings: dairy products, soy, dark-green vegetables, and whole grains. The absence of diverse vegetables, proteins, and whole grains in this dietary pattern suggests that it may diverge more from standard dietary recommendations and may indicate poor dietary balance.

A primary difference between the 2 dietary patterns may be the glycemic index of the carbohydrates consumed (34). Reductions in fasting blood glucose have been previously reported among women with gestational diabetes consuming low–glycemic index diets (35) and diets high in “complex” carbohydrates derived from vegetables, fruits, and whole grains (36). Although both dietary patterns were associated with greater GWG, only adherence to dietary pattern 2, characterized by certain high–glycemic index foods such as potatoes and other starchy vegetables, was associated with greater fasting glucose and greater newborn adiposity. These findings are also consistent with previous studies reporting inverse associations between the intake of dietary fiber or whole grains and insulin resistance (37, 38). Although our study did not directly assess the glycemic index of the maternal diet, this may be explored by future studies.

To our knowledge, there have been no previous studies describing data-driven maternal dietary patterns associated with newborn body composition. Maternal macronutrient intakes have previously been associated with newborn or infant body fat. A previous study within the Healthy Start cohort found that a greater energy intake from either fat or carbohydrates was associated with greater newborn fat mass (32). Another study conducted among pregnant women with obesity found that maternal intake of carbohydrates in late pregnancy was associated with greater newborn fat mass, but this association was not present among women in the lowest stratum of the 2-h glucose tolerance test (39). Finally, one measure of fetal adiposity (midthigh fat) was higher in fetuses of mothers with a low carbohydrate intake, but these measurements were taken during gestation and not at birth, so the results are not directly comparable to those of our study (40). By contrast, our findings suggest certain dietary patterns consumed by the mother during pregnancy may contribute to newborn size and body composition, potentially through effects on maternal fasting glucose.

An important consideration is whether differences in neonatal adiposity of the magnitude observed here will have a lasting influence on body composition and the risk of obesity. In this study, average newborn adiposity was 9%, and the 0.9% difference in adiposity we observed between the highest and lowest tertiles for dietary pattern 2 therefore represents a 10% relative increase in newborn adiposity. It is not yet known whether such changes at birth could alter a child’s susceptibility to an obesogenic environment later in life. However, in our cohort, increased adiposity at birth is a strong predictor of increased adiposity at ∼5 mo, independent of maternal and perinatal variables as well as infant feeding characteristics (41).

As in all observational studies, we are limited in our ability to make causal inferences. We find it plausible, based on other analyses in the Healthy Start study (3, 32) and in other cohorts (9, 11, 42, 43), that maternal dietary patterns may directly influence GWG and fasting glucose and that these intermediates may in turn influence newborn adiposity. We cannot exclude the possibility that the observed associations result from unmeasured confounding; however, we adjusted for many potential confounders in multiple regression models, including maternal age, race/ethnicity, education, prepregnancy BMI, parity, smoking, and physical activity during pregnancy, and the associations between the dietary patterns and newborn body composition remained statistically significant. We also adjusted for daily total energy intake and found that maternal diet composition, as defined by these dietary patterns, is related to newborn body composition independent of total energy intake in this cohort.

Maternal dietary patterns explained only 5% of the variability in the intermediates, indicating that a large fraction of GWG and fasting glucose is not determined by maternal diet but by other genetic, behavioral, and environmental factors. The percentage of variability explained is similar to that of a previous study that used reduced-rank regression to examine dietary patterns associated with cardiovascular disease risk factors in children and adolescents with type 1 diabetes (44).

A limitation of our study design is that we averaged intakes across multiple ASA24s and are therefore unable to evaluate changes in dietary composition; however, a previous study reported a minimal change in women’s dietary patterns during pregnancy (45). Women who are diagnosed with gestational diabetes may be advised to modify their diets; however, our results were robust to the exclusion of participants with diagnosed gestational diabetes. Finally, women <20 y old and women from minority racial/ethnic groups were underrepresented in this analysis; therefore, findings may not be generalizable to these groups, and the consequences of dietary composition in these groups may require further study.

Our findings indicate that maternal dietary patterns during pregnancy may be an important determinant of newborn body composition, independent of maternal obesity, total energy intake, and physical activity. Greater maternal glucose and greater newborn adiposity were observed in women who adhered to a dietary pattern characterized by the intake of eggs, potatoes and other starchy vegetables, and non-whole grains. Our findings also suggest that there may be a subtype of GWG that is metabolically favorable, at least with regard to maternal glucose concentrations during pregnancy and newborn adiposity. Women who adhered to a dietary pattern characterized by a greater intake of poultry, nuts and seeds, whole grains, cheese, and fruits had greater weight gain but no other apparent adverse metabolic effects and no increase in newborn adiposity. These results add to the evidence supporting the influence of maternal diet during pregnancy on the health of the mother and child. Continued follow-up of offspring in the Healthy Start study will allow us to examine whether increased newborn adiposity persists into childhood or whether prenatal influences may be outweighed or modified by postnatal factors, such as childhood diet and physical activity.