Associations Between Mothers' and Their Children's Weights at 4 Years of Age

Abstract

Objectives

To examine the importance of maternal weight characteristics as predictors of overweight (BMI≥85th percentile and <95th percentile) and obesity (BMI≥95th percentile) in offspring at age 4 years.

Methods

Chi-square and logistic regression analyses were conducted on a sample of 321 mother/child pairs from an earlier observational cohort study on mothers' postpartum weight retention.

Results

Maternal early pregnancy BMI and infant birth weight were each positively and significantly (p<0.05) associated with increased risk of obesity in offspring at age 4 years. A significant interaction was found between these two variables in predicting children's risk of obesity. It was driven by the high proportion of obese children among obese women who had infants weighing < 3 kg at birth. Net gestational weight gain was not associated with obesity risk in children, but was positively associated with infant birth weight among normal weight and overweight women.

Conclusions

Reducing maternal BMI in the preconception period among overweight and obese women and preventing excessive weight gain in pregnancy for all women appear to be appropriate strategies to address the childhood obesity epidemic.

Introduction

Obesity and overweight are prevalent among children in the US. In 2007–2008, 16.9% of children and adolescents age 2 through 19 years had a BMI≥95^th percentile for their age and sex and were considered obese, as were 10.4% of 2 through 5-year-old children.¹ Being obese is associated with health problems such as type 2 diabetes, hypertension, and dyslipidemia in the pediatric age group, and is also a significant risk factor for adult morbidity and mortality.²

Obesity and overweight tends to persist from the preschool years to age 12 years and beyond.³ Among children in the National Institute of Child Health and Human Development Study of Early Child Care and Youth Development, those whose BMI was ≥ 85^th percentile at one or more times from age 24 months to 54 months, were more than 5 times more likely to be overweight at 12 years of age than those whose BMI was below the 85^th percentile in their preschool years.

There is growing interest in and acceptance of the idea that there are critical or sensitive periods for the development of obesity across the life course. Several recent reviews of the literature and editorials have highlighted the role of the prenatal period in the development of childhood obesity.^4–6 The recent perspective in the New England Journal of Medicine called for the Let's Move! campaign to target Americans under 2 years of age as well as preschoolers.⁶

Maternal pre- or early-pregnancy obesity is related to increased risk of obesity (BMI≥95^th percentile) in children born to these mothers. In a sample of low-income children who participated in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) in Ohio, Whitaker⁷ showed that maternal obesity in the first trimester of pregnancy doubles the risk of obesity among children between the ages of 2 through 4 years. Using data from the National Longitudinal Survey of Youth, Salsberry and Reagan⁸ showed maternal pre-pregnancy obesity (BMI≥30) increased the risk of children aged 2 through 4 years having a BMI≥95^th percentile by about 40% compared to children whose mothers had a BMI in the normal range (18.5–24.9).

Several recent studies have shown an independent effect of pregnancy weight gain on overweight and obesity in children.^9,10 Oken et al.⁹ showed that women with adequate or excessive weight gain in pregnancy had an approximately 4-fold increased odds of having a child who had a BMI≥95^th percentile (vs. BMI<the 50^th percentile) compared to women with inadequate weight gain. Wrotniak and colleagues¹⁰ found that the odds of having a BMI≥95^th percentile in offspring at age 7 years increased 3% for every kg of gestational weight gain in participants in an older data set from the 1959–1972 National Collaborative Perinatal Project. In a study of 3-year-olds drawn from the same population as the current study, our group found the impact of pregnancy weight gain on childhood risk of BMI≥85^th percentile was modified by maternal pre-pregnancy BMI.¹¹ The effect of excess pregnancy weight gain on children's risk of BMI≥85^th percentile increased maternal BMI increased.

The purpose of the research reported here was to examine the importance of maternal weight characteristic as predictors of overweight (BMI≥85^th percentile and <95^th percentile) and obesity (BMI≥95^th percentile) in offspring at age 4 years.

Methods

Population and Sample

This study used data from the Bassett Mothers Health Project (BMHP), an observational cohort study of 622 healthy, adult women followed from early pregnancy until 2 years postpartum. Women were recruited from the population registering for prenatal care at Bassett Healthcare's network of primary care clinics in an 8-county area of rural upstate New York. All women entering care were screened for eligibility and eligible women were recruited for participation. Twenty-five percent of eligible women actively or tacitly refused participation. Additional eligibility and participation details are described elsewhere.^12,13 Of the 622 women, 593 (95.4%) delivered full-term infants with a gestational ages ≥ 37 weeks, the population of children for the current study.

The study was approved by the Institutional Review Boards of Cornell University and Bassett Healthcare.

Data Collection

For the BMHP, three methods of data collection were employed. The women completed mailed questionnaires including a food frequency at the first or second trimester of pregnancy, 6 months postpartum, 1 and 2 years postpartum. Shortly after delivery, the women's obstetrical records were audited and data entered directly into a computer. In addition, women's weights and heights were measured following study protocols by health care providers at antenatal visits and at 1 and 2 years postpartum. The obstetrical record was the primary source for the maternal data included in this article.

In 2009 and 2010, the medical records of the infants were located through the Bassett Healthcare system and audited for anthropometric and health related information. Of the 593 children, 321 had measured heights and weights between the ages 42 and 54 months (3.5 to 4.5 years), the age range of interest for this article.

Measures and Variables

The specific methods and instruments used in the BMHP are described in detail in a previous publication¹⁴ and described briefly here. For mothers, BMI at several time points was used as a continuous variable and also as body weight categories. Using measured weights from early in the first trimester of pregnancy, 1 and 2 years postpartum to calculate BMI, women's body weights were classified as low, normal, high, or obese using the Institute of Medicine (IOM) BMI categories: Underweight (<18.5), normal (18.5 to <25.0), overweight (25.0 to <30.0), and obese (≥30.0).¹⁵ For some of the analyses, underweight and normal BMI women were combined into one group (not overweight) and the overweight and obese BMI women were combined into a single group (overweight). Gestational weight gain was expressed as gaining less weight than recommended (inadequate), the recommended amount (appropriate), or more than recommended (excessive) according to the IOM BMI categories and gestational weight gain guidelines (15). Net gestational weight gain was used as a continuous variable and was calculated by subtracting infant birth weight from total gestational weight gain. Household income level was categorized as low income (≤185% of the federal poverty line) or not low income (>185% of the federal poverty line). Variables considered as potentially confounding variables in the relationship between maternal weight characteristics and child overweight included: smoking during pregnancy (yes/no); parity (nulliparous/parous); and infant birth weight (expressed in grams). The children's BMI percentile was determined by a SAS program available on the Centers for Disease Control and Prevention (CDC) growth charts web page and based on the CDC growth reference for 2000.¹⁶ Children with a BMI percentile≥the 85^th and < 95^th sex-specific percentile for age were defined as overweight and those ≥ 95^th sex-specific percentile for age were defined as obese.

Data Analysis

Chi square analyses were conducted to compare children included in the analysis sample to those not included in order to identify any factors that might be different. Chi square analyses were also conducted to compare the proportion of children who were normal weight, overweight and obese across categories within maternal and infant predictors of child weight. Logistic regression models were then developed for predicting overweight and obesity in children controlling for potentially confounding factors. Because some of the continuous predictors were correlated, such as net gestational weight gain and BMI, these maternal weight factors were centered at the mean to reduce the effect of collinearity. In building the logistic regression models, the maternal factors of interest (BMI and net gestational weight gain) and potentially confounding factors (infant birth weight, household income, maternal parity, and smoking) were included, controlling for gestational age and sex of the infant. All two-way interactions in the full model were assessed and included in the final model if the p-value associated with the factor was ≤ 0.05.

Results

Description of the Sample

The characteristics of the population-based sample and the sample used in the analyses are shown in Table 1. While those mother-infant pairs included in the analysis sample did not differ significantly (p<0.05) from those who were not included, differences approached statistical significance for several variables. The analysis sample had a lower proportion of mothers who smoked during pregnancy (16.5%, p=0.069), a higher proportion of mothers who were having their first child (44.5%, p=0.082), and a higher proportion of female infants (51.4%, p=0.071). Most importantly, the samples did not differ on the proportion of mothers who were overweight in early pregnancy (50%) or gained more than the recommended amount of weight in pregnancy (48%). Thus the analysis sample is a reasonable one for investigating the relationship between maternal and child weight characteristics.

Table 1.

Characteristics of population and analysis sample

CHARACTERISTIC	POPULATION SAMPLE N=593		ANALYSIS SAMPLE N=321		COMPARISON OF ANALYSIS SAMPLE TO THOSE NOT INCLUDED
	N	PERCENT	N	PERCENT	p-VALUE**
Maternal Characteristics
Overweight in early pregnancy*					0.744
Yes	294	49.60%	161	50.20%
No	299	50.40%	160	49.80%
Pregnancy weight gain >recommended*					0.935
Yes	282	47.60%	154	48.00%
No	311	52.40%	167	52.00%
Low income					0.184
Yes	259	43.70%	132	41.10%
No	325	54.80%	187	58.30%
Missing	9	1.50%	2	0.60%
Smoking during pregnancy					0.069
Yes	114	19.20%	53	16.50%
No	479	80.80%	268	83.50%
Nulliparous					0.082
Yes	242	40.80%	143	44.50%
No	350	59.00%	178	55.50%
Missing	1	0.20%
Baby Characteristics
Sex					0.071
Male	309	52.10%	156	48.60%
Female	280	47.20%	165	51.40%
Missing	4	0.70%
Birth order					0.124
1	242	40.80%	143	44.50%
2	202	34.10%	109	34.00%
3 or later	148	25.00%	69	21.50%
Missing	1	0.20%
Birth weight >4000g					0.435
Yes	136	22.90%	68	21.20%
No	457	77.10%	253	78.80%
Birth weight (g)
Mean (SD)	3623.8	(510.5)	3590.95	(508.7)

^*Based on new CDC BMI and 2010 IOM Recommendations for GWG.

^**p-value based on comparison between analysis sample and those not included in the analysis from the original sample.

Associations between Maternal Weight Characteristics and Weight of Children

Overall, 32.1% of the preschoolers were ≥ 85^th percentile, 17.4% (N=56) were ≥ 85^th and < 95^th percentile and classified as overweight, and 14.6% (N=47) were ≥ 95^th percentile and classified as obese. As shown in Table 2, maternal early pregnancy weight category was significantly (p<0.0001) associated with children's weight status. Thirty-six percent of the children of obese mothers were obese, whereas only 7.8% and 10.3% of children of mothers of normal weight and overweight, respectively, were obese. Further analyses (data not shown) indicate children whose mothers were overweight (overweight and obese maternal BMI combined) in early pregnancy were significantly more likely to be ≥ 85^th and 95^th percentiles at 4 years than children whose mothers were not overweight (underweight and normal BMI combined) in early pregnancy (39.6% vs. 26.4%; p<0.05 and 25.2% vs. 6.6%; p<0.001, respectively). There was no significant difference in the proportion of preschoolers who were overweight but not obese (BMI≥85^th percentile and <95^th percentile) by maternal early pregnancy weight status (19.8% vs. 14.4%).

Table 2.

Proportion of children who are normal weight, overweight and obese by maternal and infant weight characteristics

	CHILDREN'S WEIGHT STATUS
CHARACTERISTIC	% Normal (N=218)	% Overweight (N=56)	% Obese (N=47)
Maternal Early Pregnancy Weight*
Underweight	79 (23)	21 (6)	0 (0)
Normal weight	73 (111)	20 (30)	8 (12)
Overweight	72 (42)	17 (10)	10 (6)
Obese	52 (42)	12 (10)	36 (29)
Infant Birth Weight
< 3 kg	67 (22)	15 (5)	18 (6)
≥ 3<3.5 kg	78 (84)	12 (13)	10 (11)
≥ 3.5<4 kg	63 (71)	21 (23)	16 (18)
≥ 4 kg	60 (41)	22 (15)	18 (12)
Gestational Weight Gain
Inadequate	68 (42)	16 (10)	16 (10)
Appropriate	71 (92)	18 (23)	11 (14)
Excessive	65 (84)	18 (23)	18 (23)
Maternal Weight Change from 1 to 2 Years Postpartum (N=245)
Decreased≥5 pounds	73 (49)	8 (5)	19 (13)
Stayed within 5 pounds	66 (78)	23 (27)	12 (14)
Increased≥5 pounds	70 (41)	15 (9)	15 (9)

^*p<0.0001.

Maternal weight at 2 years postpartum was highly correlated with early pregnancy weight (r=0.94, p<0.001). However, weight change between 1 and 2 years postpartum was not correlated with early pregnancy weight so this variable was used to reflect maternal weight later in the children's lives. While overall, this variable was not significantly associated with children's weight, children whose mothers had lost weight were significantly less likely (p<0.05) to be overweight than those of mothers who were weight stable or gained weight (Table 2). Neither infant birth weight or gestational weight gain were significantly associated with children's weights in the unadjusted analyses.

Table 3 shows the results of the multivariate analyses of the association between maternal weight characteristics (as continuous variables) and the children's risk of overweight and obesity, controlling for confounding factors. Similar to the unadjusted analyses with categorical variables, maternal BMI was a significant (p=0.001) positive predictor of offspring risk of being obese, but was not a significant (p=0.344) predictor of a child's risk of being overweight. In the multivariate analyses, infant birth weight was a significant positive predictor of both child weight outcomes, increasing risk by approximately 300% for each kg of birth weight (overweight p=0.008; obesity p=0.015. There was a significant (p=0.004) negative interaction between maternal early pregnancy BMI and infant birth weight in predicting obesity, but not for overweight. No other variables in the multivariate model, including net gestational weight gain, were statistically significant. Smoking during pregnancy approached statistical significance (p=0.077) in the expected direction in the overweight model.

Table 3.

Multivariate logistic regression models showing predictors of childhood risk of overweight (BMI≥85th and < 95th percentile) and obesity (BMI≥95th percentile) at age 4 years

	OVERWEIGHT MODEL*		OBESITY MODEL*
	R2=7%		R2=28%
	OR	p-VALUE	OR	p-VALUE
Maternal (m) BMI	1.208	.344	2.268	.001
Net Gestational Weight Gain	1.001	.956	1.011	.459
Birth Weight (in kg)	2.880	.008	3.216	.015
mBMI×Birth Weight	0.946	.292	0.833	.004

^*Gestational age and parity are included in the models as control variables and are significant. Household income level, maternal smoking in pregnancy, and sex of the child are included and are not statistically significant.

OR, model estimated odds ratio per unit change; R², = Max-rescaled R-square statistic;

P-value, p-value for the test that the OR=1

Additional analyses (data not shown) demonstrate that maternal net gestational weight gain was positively and significantly related to infant birth weight for normal weight (r=0.251, p=0.002) and overweight mothers (r=0.295, p=0.024). A significant relationship was not found in underweight (r=0.130, p=0.502) and obese (r=0.017, p=0.884) mothers.

The Interaction of Maternal BMI and Birth Weight on Offspring Obesity Risk

Figure 1 illustrates the significant negative interaction shown in Table 3 between maternal early pregnancy weight and infant birth weight in predicting obesity risk in offspring (OR=0.833, p=0.004). The interaction is driven by the large proportion of children who were obese (67%) and were born weighing < 3 kg to mothers who were obese. This interaction should be interpreted with caution because the 67% represents 6 of 9 children, so the sample size in this cell is small. In all other maternal weight groups, the association between birth weight and obesity risk is positive, as birth weight increases so does obesity risk. Also, no matter what the birth weight, no children born to underweight mothers were obese and no < 3 kg babies born to normal weight and overweight mothers were obese at 4 years of age.

Figure 1.

Proportion of children who were obese within each maternal early pregnancy and infant birth weight category.

Discussion

Higher maternal early pregnancy BMI was consistently associated with increased likelihood that the offspring would be obese at age 4 years in several different approaches to data analysis. Infant birth weight, as a continuous variable, was a significant positive predictor of both overweight and obesity in children in analyses controlling for confounding factors. In these analyses, we found a significant interaction between maternal BMI and birth weight in predicting childhood obesity, but urge caution in interpreting this interaction. It is only among obese mothers that the relationship of birth weight to the risk of obesity is in a negative direction. In bivariate analyses, net gestational weight gain was positively associated with birth weight in normal and overweight BMI women. Taken together, these findings suggest than any influence of gestational weight gain on child risk of overweight and obesity is likely mediated through its effect on birth weight. Our finding of higher birth weight associated with increased child risk of overweight and obesity is consistent with the literature.^5,7,8,17

The results of this study support the earlier findings of Whitaker⁷ and Salsberry and Regan⁸ on a positive relationship between maternal BMI and obesity risk in offspring, as well as, more recent findings of Li and colleagues¹⁷ on early and late onset overweight. We did not replicate Oken et al.'s⁹ and Wrotniak et al.'s¹⁰ findings of an independent effect of excessive gestational weight gain on risk of overweight and obesity in offspring. This is very likely due to the differences in the maternal weight characteristics of the samples. In Oken et al.'s⁹ and Wrotniak et al.'s¹⁰ samples, only 9.2% and 16% of the women were obese, respectively, while 26.6% were obese in our sample, using the 1990 IOM cut-offs for BMI to define the weight groups. The respective proportions for overweight were 17.7%, 29.0% , and 43.0%. If the effect of gestational weight gain on child obesity risk is mediated through birth weight and the impact of birth weight on child weight decreases with increased maternal BMI, then one would expect to see less of a relationship between gestational weight gain and child obesity outcomes in samples with a large proportion of overweight and obese women.

The magnitude of the childhood obesity epidemic calls for multiple interventions on many fronts. An implication of this study is that maternal weight characteristics in pregnancy cannot be ignored in addressing the epidemic. The American College of Obstetricians and Gynecologists recommends that obese women should undertake a weight reduction program before pregnancy.¹⁸ This is a laudable goal, but may not be realistic for all women. In a recent review, intervention during pregnancy to reduce the prevalence of excessive gestational weight gain through clinical and patient education interventions was shown to successful in some of the studies.¹⁹

Conclusions

Maternal early pregnancy BMI is a significant and important influence on risk of obesity in children during the preschool years. Infant birth weight is also an important predictor. Higher levels of net gestational weight gain increase the risk of overweight and obesity in childhood likely through their effect on birth weight, among normal weight and overweight women. Interventions to reduce maternal pre-pregnancy BMI in overweight and obese women and excessive gestational weight gain in all women should be part of the overall strategy to prevent childhood obesity.

Acknowledgments

The funding for the Bassett Mothers Health Project was provided by the National Institute of Child Health and Human Development (Grant No. HD 29549). Additional funding was provided by NIH Grant No. DK69448. The doctoral students are supported by a USDA National Needs Graduate Fellowship Competitive Grant No. 2008-38420-04825 from the National Institute of Food and Agriculture (Demment) and a NIH Ruth L. Kirschstein National Research Service Award (NRSA) Institutional Research Training Grant No. 5-T32-DK-0007158 (Carling). In addition, the authors acknowledge the assistance with the child chart audit process provided by Drs. John May, Chris Kjolhede, and medical records staff of Bassett Healthcare.

Author Disclosure

The author has no conflicts of interest to disclose.