Main

Left ventricular hypertrophy is a strong and independent risk factor of cardiovascular morbidity and mortality (1). Left ventricular mass (LVM) tracks from childhood to adulthood, meaning that, relative to body size, the size of the heart of a subject remains in a given rank order over time compared with other subjects (2). The human heart has its highest growth rates in fetal and early postnatal life (3). After this period, increase in cardiac muscle mass is mainly due to enlargement of preexisting cells (4). This rapid growth in fetal and early postnatal life suggests that early life factors may permanently affect cardiac growth and development (3).

The developmental origins of health and disease hypothesis postulates that adverse environmental exposures in fetal and early postnatal life lead to adaptations that permanently program the fetus' structure, physiology, and metabolism (5,6). This hypothesis is supported by both animal studies and epidemiologic studies, which have consistently shown associations of fetal and early postnatal growth characteristics with development of cardiovascular disease and its risk factors (7,8). However, fetal cardiovascular adaptations may be present without overt changes in fetal and early postnatal growth characteristics. Maternal anthropometrics during pregnancy are related to their nutritional and health status and thereby to fetal environmental exposures (9,10). Several studies have shown that maternal anthropometric factors such as body mass index (BMI), weight gain during pregnancy, and insulin resistance are associated with fetal growth characteristics and adverse pregnancy outcomes (11–15). Recently, it has also been suggested that both maternal gestational diabetes and macrosomia in children of mothers without overt gestational diabetes were associated with newborn LVM (16). Thus, maternal nutritional and health status during pregnancy may lead to an adverse fetal environment and might affect cardiac growth and development.

We hypothesized that maternal anthropometrics during pregnancy affects cardiac growth and development in early postnatal life. Therefore, we examined in a population-based, prospective cohort study whether maternal height, weight, and BMI in various periods of pregnancy and maternal weight gain during pregnancy are associated with LVM growth between 6 wk and 6 mo.

PATIENTS AND METHODS

Design.

This study was embedded in the Generation R Study, a population-based prospective cohort study from fetal life until young adulthood. This study is designed to identify early environmental and genetic determinants of growth, development, and health from fetal life until young adulthood and has been described previously in detail (17,18). In total, the cohort includes 9778 mothers and their children living in Rotterdam, the Netherlands. A vast majority of mothers were enrolled in the first trimester of pregnancy (18). Assessments in pregnancy included physical examinations, fetal ultrasounds, biologic samples, and questionnaires and were planned in early (gestational age <18 wk), mid- (gestational age 18–25 wk), and late pregnancy (gestational age >25 wk) to collect information about fetal growth and its main determinants. In this cohort, additional, more-detailed assessments of fetal and postnatal growth and development are conducted in a subgroup of 1232 Dutch mothers and their children from late pregnancy. This subgroup is referred to as the Generation R Focus cohort. This subgroup is ethnic homogeneous to exclude possible confounding or effect modification by ethnicity. Dutch ethnicity was defined as having two parents and four grandparents born in the Netherlands (18). Between February 2003 and April 2005, all mothers participating in the Generation R Study and pregnant of children who met this criterion were approached for these additional measurements in late pregnancy (gestational age 30 wk). No other exclusion criteria were used. Of all approached women, 80% agreed to participate in this subgroup study. In this subgroup, follow-up postnatal echocardiograms were performed in infants at the ages of 6 wk and 6 mo. The study has been approved by the Medical Ethics Committee of the Erasmus Medical Center, Rotterdam. Written informed consent was obtained from all participants.

Population for analysis.

In total, 1232 women were enrolled in the Generation R Focus Study at a gestational age of 30 wk. The present analysis was limited to singleton live births (n = 1,215). Twin pregnancies (n = 15) and pregnancies leading to intrauterine or perinatal death (n = 2) were excluded from the analysis. None of the children in this analysis had congenital heart disease. Of the singleton live births, 76% (n = 923) participated in the postnatal assessment at 6 wk of age and 74% (n = 903) participated in the postnatal assessments at 6 mo of age. Echocardiograms were performed in 85% (n = 791) of these infants at 6 wk and in 87% (n = 785) of these infants at 6 mo of age. Missing echocardiograms were mainly due to crying behavior or unavailability of equipment or echocardiographer.

Maternal anthropometrics.

Maternal anthropometrics were measured in one of the research centers at the visits in early (gestational age <18 wk), mid- (gestational age 18–25 wk), and late pregnancy (gestational age >25 wk). We measured weight (kg) and height (cm) and we calculated maternal body mass index (kg/m2) for each pregnancy period. The gestational age medians (95% range) for these assessments were 12.6 (9.6–16.9) wk, 20.4 (18.6–22.5) wk, and 30.2 (28.5–32.5) wk, respectively. Information about maternal weight just before pregnancy was obtained from questionnaires. Self-reported weight just before pregnancy was highly correlated to measured weight at intake (r = 0.97, p value <0.01). Weight gain was defined as the difference between weight before pregnancy and in late pregnancy (gestational age 30 wk). This is a measure of first two-trimester weight gain.

Left cardiac structure.

Two-dimensional M-mode was performed on Kretz Voluson 530D equipment in the children at 6 wk and 6 mo of age. The examination was carried out in a quiet room with the baby quietly awake in a supine position. One echocardiographer performed the vast majority (86%) of these measurements. Two other echocardiographers performed the other measurements. In a parasternal long-axis view, left ventricular end diastolic diameter (LVEDD), interventricular septal thickness (IVST), and left ventricular posterior wall thickness (LVPWT) were measured using methods recommended by the American Society of Echocardiography (19). LVM was computed by use of the formula derived by Devereux et al. (20):

LVM = 0.80 * 1.04 ((IVST + LVEDD + LVPWT)3 – (LVEDD)3) + 0.6.

The intraobserver intraclass correlation coefficients for the echocardiographic measurements in the present study were similar as previously demonstrated in large scale multicenter studies in young children (21,22).

Covariates.

All anthropometrics in the infants were measured without clothes at the same visits as the echocardiograms at 6 wk and 6 mo of age. Weight was measured to the nearest gram using electronic scales. Length was measured in supine position to the nearest millimeter using a neonatometer. Birth weight, date of birth, and gender were obtained from midwife and hospital registries.

Data analysis.

Associations of maternal anthropometrics with repeatedly measured LVM at 6 wk and 6 mo of age were first assessed using repeated measures regression analysis using the Proc Mixed module of SAS (23). This is a regression technique that takes the correlation of multiple measurements within one subject into account and is used to examine response trends over time (23). Maternal anthropometrics included weight and body mass index measured in early, mid-, and late pregnancy and weight gain during pregnancy. Based on tertiles of weight gain until late pregnancy within our study population, we constructed three categories for increase in weight (<6.5 kg, 6.5–9.0 kg, and > 9.0 kg). Weight gain during pregnancy was included in the model as an interaction term with age (p value <0.05). This model enables us to assess both the time independent and time dependent effects of weight gain on LVM. The model can be written as:

LVM = β0 + β1 * increase in weight + β2 * age + β3 * increase in weight * age.

In this model, the term including “β0 + β1” reflects the intercept and the term including “β2” reflects the slope of growth in LVM per week. The term including “β3” reflects the differences in growth of LVM between the different categories of increase in weight. First, we examined the association of maternal weight, BMI, and weight gain in pregnancy with LVM in the offspring adjusted for gender (Model I). All models were additionally adjusted for current weight and current length of the children (Model II). Models with maternal weight gain as independent variable were additionally adjusted for maternal weight and height just before pregnancy (Model II).

Subsequently, we performed multiple linear regression models to study the associations of maternal anthropometrics with LVM at the ages of 6 wk and 6 mo separately. These models were first adjusted for gender and age at the visit (Model A) and additionally for current weight and current length of the children (Model B). All measures of association are presented with their 95% confidence intervals (CI). Statistical analyses were performed using the Statistical Analysis System version 8.2 (SAS, Stata Corporation, College Station, TX), including the Proc Mixed module for unbalanced repeated measurements and the Statistical Package of Social Sciences version 11.0 for Windows (SPSS Inc, Chicago, IL).

RESULTS

Characteristics of infants who participated in the postnatal echocardiogram studies and their mothers are presented in Table 1. The percentage of boys was 52%. Birth weight was larger in boys than in girls. The overall median age in infants at their 6 wk visit was 6.4 (95% range: 4.4–12.5) wk and at their 6 mo visit 27.3 (95% range: 23.7–35.6) weeks. Weight, length, and LVM were somewhat smaller in girls than in boys. Infants who had a postnatal echocardiogram (n = 791) did not differ from the postnatal nonresponders (n = 424) in maternal weight [difference 0.6 (95% CI: –1.1, 2.3) kg] and maternal BMI [difference 0.2 (95% CI: –0.4, 0.7) kg/m2] just before pregnancy. Figure 1 shows the correlations between infant characteristics and LVM.

Table 1 Subject characteristics
Full size table
Figure 1
figure 1

Correlations of individual measurements of infant weight and length with LVM at the ages of 6 wk and 6 mo. (A) Correlation weight and LVM at the age of 6 wk, r = 0.30, p value <0.01. (B) Correlation length and LVM at the age of 6 wk, r = 0.29, p value <0.01. (C) Correlation weight and LVM at the age of 6 mo, r = 0.35, p value <0.01. (D) Correlation length and LVM at the age of 6 mo, r = 0.33, p value <0.01.

Full size image

The associations between maternal anthropometrics during pregnancy and longitudinal measured LVM in infancy are presented in Table 2. No associations were found of maternal weight and BMI in early, mid-, and late pregnancy with growth of LVM. Compared with mothers with the lowest increase in weight during pregnancy, growth of LVM was 0.28 (95% CI: –0.17, 0.72) g/wk lower in mothers with moderate increase in weight during pregnancy and 0.46 (95% CI: 0.02, 0.90) g/wk higher in mothers with the highest increase in weight during pregnancy.

Table 2 Maternal anthropometrics and tertiles of weight gain in pregnancy and longitudinal measured growth in LVM
Full size table

Table 3 gives the associations between maternal weight gain during pregnancy with LVM at the ages of 6 wk and 6 mo separately. We did not find an association of maternal weight gain during pregnancy with LVM at 6 wk of age (Table 3). However, weight gain during pregnancy was positively associated with postnatal LVM at the age of 6 mo. For each kg increase in weight during pregnancy, LVM at the age of 6 mo increased by 0.08 (95% CI: 0.02, 0.15) grams. This means an increase of 0.47%. The associations did not materially change after adjustment for maternal weight and height just before pregnancy, and gender, age, weight and length of the child (Table 3).

Table 3 Maternal weight gain during pregnancy and LVM at 6 wk and 6 mo
Full size table

DISCUSSION

Our study demonstrates for the first time that higher maternal weight gain during pregnancy is associated with a higher increase in growth of LVM during the first 6 mo of life. These findings suggest that maternal anthropometrics during pregnancy may have consequences for LVM in the offspring.

Methodological considerations.

Postnatal follow-up echocardiograms were successfully obtained in 65% of all singleton live births of the mothers participating in the Generation R Focus Study. Maternal anthropometrics in pregnancy were similar between infants who had a postnatal echocardiogram and the postnatal nonresponders. The observed effect estimates would be biased if the associations of maternal anthropometrics with LVM differ between those included and not included in the present analyses. This seems unlikely. The missing values of LVM among infants who did participate in postnatal visits were mainly due to crying behavior or unavailability of equipment or echocardiographer and are unlikely to lead to biased results.

Weight gain was partly based on self-reported weights, which is known to introduce measurements error. Based on an assessment of validity of self-reported weight in the US Third National Health System, women in this age group may systematically underestimate their weights (24). Inasmuch as self-reported weight just before pregnancy was highly correlated (r = 0.97) to measured weight at intake, we do not think that this would lead to biased results.

In the present analysis, we defined maternal weight gain as the difference between weight in late pregnancy and weight just before pregnancy. In fact, this is the increase in weight in the first two trimesters of pregnancy. Ideally, weight gain during pregnancy should be defined as the difference between the highest weight in pregnancy and the weight just before pregnancy. Since the largest differences in maternal weight gain during pregnancy are expected to appear in late pregnancy, our associations between weight gain and LVM may be underestimated.

Maternal anthropometrics and LVM.

No associations were found of maternal weight and BMI in early, mid-, and late pregnancy with LVM in infancy. We found that mothers who increase more in weight during pregnancy have children with an increased growth of the LVM in the first 6 mo, resulting in larger LVM at the age of 6 mo. These associations were independent of maternal weight just before pregnancy.

Anthropometrics during pregnancy reflect maternal nutritional and health status and may be measures of fetal environmental exposures (10). Increased maternal weight gain during pregnancy may be caused by several factors including nutrition, obesity, fluid retention, and lower physical activity levels (25). Furthermore, increased maternal weight gain during pregnancy is associated with various pregnancy complications including insulin resistance, gestational diabetes, and pregnancy-induced hypertension. Therefore, several underlying biologic pathways may explain the associations between maternal anthropometrics in pregnancy and LVM in infancy. The usual increase in insulin resistance seen in late pregnancy is higher in mothers who show a marked increase in weight during pregnancy (26). Insulin resistance causes higher glucose levels in the mother. Since glucose passes the placenta to the child, fetal insulin production will be increased. Insulin is the main fetal growth factor and may stimulate fetal body and heart growth (27). Offspring of mothers with gestational diabetes have increased cardiac sizes in the neonatal phase (16). However, this larger cardiac size is usually not present anymore at the age of 6 mo (28). In our study, data on subclinical insulin resistance or increased glucose levels during pregnancy were not available. The number of pregnant women with doctor-diagnosed gestational diabetes in our study was too small (n = 45) to assess the effect on LVM. Another biologic pathway may directly be related to the nutrition intake of the mother during pregnancy (29,30). Excessive fetal exposure to energy sources may lead to rapid growth and cardiac enlargement. Further studies, in which measures of insulin resistance such as insulin, glucose, or HbA1c levels and maternal nutrition in pregnancy have been measured, are needed to identify the underlying causal pathways.

The associations could also be explained by postnatal factors related to postnatal growth, nutrition, blood pressure development, and genetic susceptibility (31–33). We showed that the effect of weight gain on LVM was still present after adjustment for current weight. Thus, we do not think that postnatal anthropometrics explain the associations. However, previous studies showed that in adolescents, increase in BMI from early childhood, rather than current BMI is associated with LVM (32,33). Thus, postnatal growth patterns or their determinants, which are also associated with a maternal weight gain in pregnancy, may explain part of these associations. Further follow up studies in our and other cohorts may reveal these underlying mechanisms.

The effect estimates shown in our study are unlikely to be of any clinical significance in infancy. However, our results suggest an etiological role of maternal anthropometrics during pregnancy in LVM development. Increased LVM may precede the development of left ventricular hypertrophy and increased blood pressure. Left ventricular hypertrophy and hypertension are risk factors for cardiovascular morbidity and mortality (1). Whether and to what extent the effects of maternal anthropometrics on LVM persists in childhood and in adulthood needs to be further studied.

Study implications.

Our findings demonstrated that maternal anthropometrics, as measure of fetal environment, are associated with LVM in infancy. The effect sizes seem to be too small for clinical implications. However, these results are of interest from a developmental perspective and add new support for the developmental origins of health and disease hypothesis. Further studies examining the underlying biologic pathways and the long-term consequences are needed.