Obesity is a major public health concern in the United States and worldwide today.1 Obesity plays an important role in the rising rates of chronic health conditions in both children and adults and contributes to increasing health care costs.2 Although high rates of obesity were first observed among adults, a dramatic rise in the prevalence of childhood obesity has emerged as well.3–5 Estimates from the National Health and Nutrition Examination Survey indicate that 12.1% of 2- to 5-year-old, 18.0% of 6- to 11-year-old, and 18.4% of 12- to 19-year-old children in the United States were obese in 2009–2010.6 Young overweight and obese children are at increased risk of being overweight and obese as adolescents and adults.7,8 The U.S. Centers for Disease Control and Prevention (CDC) forecasts that the current generation of children will be the first to have a shorter life expectancy than their parents, attributable in part to obesity.
Current research suggests a mother’s influence on the risk of obesity in her offspring begins early, the result of her physical health before and during pregnancy.9 Prepregnancy obesity, gestational weight gain, gestational diabetes mellitus (GDM), and tobacco use during pregnancy have each been implicated as risk factors for obesity in the offspring and targets for intervention during the prenatal period.10 It is uncertain, however, whether maternal factors during pregnancy have a significant relation to young children’s body mass index (BMI) independent of the mother’s preconception health. Prior epidemiologic research has been limited by the paucity of information about maternal health risks before pregnancy and the pregnancy course included in large observational studies of children’s growth. A better understanding of the role of early factors is needed to determine the contribution of maternal health to a child’s future obesity and to guide the development of effective interventions.
The objective of the study reported here was to examine the independent contribution of risk factors developing during pregnancy to subsequent risk of obesity in young children. Data were obtained from a contemporary cohort constructed using obstetric and pediatric records for mother–child dyads. These data provided a unique opportunity to explore the influence of maternal health for a large and diverse sample of mothers and their children.
MATERIAL AND METHODS
The study sample was drawn from a cohort of 4,852 mother–child dyads, the Delaware Mother Baby Cohort based on data derived from electronic medical records of mothers and their children. Obstetric labor and delivery records of all singleton births from 2004 to 2007 at Christiana Care Health System were linked to the electronic medical records of children who attended a well-child visit at one of the Nemours primary care practices. This study was approved by the Christiana Care Health System Institutional Review Board in coordination with the Nemours Institutional Review Board.
All healthy children from the Delaware Mother Baby Cohort were included in the study. Children with a diagnosis of a birth defect (n=102) or other significant chronic disease (such as cancer, severe neurologic disability, chronic gastrointestinal disease, severe or chronic infectious disease, or a metabolic disorder) noted in the problem list in the electronic medical record (n=76) were excluded. Children were also excluded if they did not have a clinical visit between 45 and 57 months of age (n=1,369). An additional three children who had BMI z scores of more than 5 standard deviations from 0 were excluded. Mothers with missing data for prepregnancy BMI or gestational weight gain were excluded from the analyses where these data where required; 18 (0.5%) had missing data for maternal prepregnancy BMI and 57 (1.7%) a missing value needed to calculate net gestational weight gain.
The final sample for analysis included 3,302 children: 68% of the original cohort and 80% of children who had reached at least 4 years 9 months of age at the date of records extraction. The sample for analysis included a smaller prevalence of chronic hypertension, diabetes mellitus, tobacco use, or “other” race than among the original cohort of mothers. There were no significant differences in mean maternal BMI or gestational weight gain or in mean birth weight or gestational age of the children with and without follow-up data at 4 years of age.
Data for the mother and newborns were entered by nurses into the obstetric electronic medical record of women on admission to and during their stay in labor and delivery. The sources of maternal information were from admitting physician as well as the prenatal summary sheet provided by the primary obstetrician. Clinical diagnoses were determined by the patient’s physician.11,12 Data for the children were obtained from their records at Nemours practices where information had been entered by nurses and physicians during the course of clinical care.
Mother’s prepregnancy weight, weight at delivery, and height were primarily self-reported. Maternal BMI was calculated as weight (kg)/[height (m)]2 and categorized as less than 18.5 (underweight), 18.5–24.9 (optimal), 25–29.9 (overweight), 30–39.9 (obese), and 40+ (severely obese) according to World Health Organization and National Heart, Lung, and Blood Institute groupings.13 Gestational weight gain was measured using several definitions and approaches to first explore the underlying relations as well as understand outcomes related to the Institute of Medicine gestational weight gain guidelines. Total gestational weight gain was defined as the difference between the mother’s weight at delivery and prepregnancy weight in kilograms. Net gestational weight gain in kilograms was calculated as the gestational weight gain subtracting birth weight. Because gestational weight gain increases with gestational age, an adjusted net gestational weight gain was derived by controlling for gestational age using the residuals remaining after linear regression. The adjusted net gestational weight gain was included in the multivariable linear regressions. Women delivering at term (37 or more weeks of gestation completed) were categorized based on their prepregnancy BMI category using the 2009 Institute of Medicine gestational weight gain guidelines as having greater than (excess), less than (inadequate), or recommended weight gain.14
Tobacco use during pregnancy was considered present if a woman reported smoking during any trimester of pregnancy. Race and ethnicity were self-identified and categorized as white, African American, Hispanic, and “other”; white was the referent group. Insurance was defined as private insurance compared with Medicaid or uninsured as the referent group. Parity was categorized as nulliparous with multiparous used as the referent group. Maternal age at delivery was in years.
Children’s anthropometric measures between 45 and 57 months were obtained from the pediatric clinical record and included height, weight, and calculated BMI. The vast majority (98.5%) of anthropometric measures for children were obtained during a well-child visit. Electronic medical records have been used successfully by several investigators to study child growth and development; and methods of data cleaning have been validated.15 Measurements were obtained using standard pediatric office procedures during the course of clinical care and entered into the pediatric electronic medical record. Body mass index z score for sex and age in months was calculated using the 2000 CDC Growth Charts.16 The z score is the number of standard deviations the child’s BMI falls from the mean in the normative population for sex and age in months. Obesity at 4 years of age was defined as a BMI z score 95th percentile or greater and overweight or obese as a BMI z score 85th percentile or greater.
The primary dependent variable was BMI z score at 4 years of age. Although BMI is a surrogate measure of adiposity, it was chosen because it is easily measured and commonly used in clinical practice as well in the standard international definition of obesity and has known association with obesity-related morbidities. Maternal prenatal health risks included measures of gestational weight gain, GDM, gestational hypertension and preeclampsia, and tobacco use during pregnancy. Prepregnancy BMI, pre-existing diabetes mellitus, chronic hypertension, and maternal demographic characteristics were examined as measures of maternal preconception health.
Univariable relations between each maternal factor and BMI z score were examined using simple linear regression and analysis of variance with Bonferroni post hoc estimation. Alpha <.05 defined statistical significance. Multivariable linear regression was used to estimate the relation of maternal health risks with BMI z score at 4 adjusting for covariates. For analyses that included the Institute of Medicine recommendations for gestational weight gain, mother–child dyads with births occurring before term (less than 37 completed weeks of gestation) were excluded. Regression models were estimated first for the most proximal maternal health risks and then for the most distal. Model 1 contained pregnancy-related factors including gestational weight gain, GDM, gestational hypertension, tobacco use during pregnancy, and parity. Model 2 included maternal prepregnancy BMI and diagnosis of diabetes mellitus and chronic hypertension. Model 3 added the mothers’ demographic characteristics including race and ethnicity, insurance type, marital status, and maternal age at delivery. All models were adjusted for the child’s age in months at the time of weight and height measurement and maternal height.
To explore the potential effect of underestimation of gestational weight gain by mothers, a sensitivity analysis was conducted to estimate the difference in the coefficients if the gestational weight gain was 10% or 20% greater for each woman and to ensure that misclassification of gestational weight gain group did not change the results. There were no significant differences in the estimates of the relation between gestational weight gain and BMI z score; therefore, these results are not reported. The presence of significant two-way interactions between gestational weight gain and other covariates was excluded using stratified linear regression as well as interaction terms.
Because some mothers contributed more than one birth to the cohort, hierarchical modeling was used to account for the shared characteristics of mothers of siblings. All analyses were completed using Stata/IC 10.1.
Characteristics of the mothers are shown in Table 1. The sample was diverse with 39% of mothers white, 46% African American, 11% of Hispanic ethnicity, and 3% “other” race. The “other” group included 67% Asian, 0.1% Native American, and the balance unknown or mixed race. The median age at delivery was 26 years (range 11–47 years). The mean BMI of the mothers differed significantly by race and ethnicity (P<.001): the “other” group was the lightest (24.7) followed by white (25.6) and Hispanic mothers (26.3); African American women had the highest prepregnancy BMI (27.8). The prevalence of prepregnancy obesity among white mothers was 16.2%; among African American mothers, 24.5%; and among mothers of Hispanic ethnicity, 19.7%. Net gestational weight gain also differed by race: white mothers gained the most weight (11.9 kg) and significantly more than African American mothers (10.8 kg, P<.001) and mothers of Hispanic ethnicity (10.2 kg, P<.001).
The children in the cohort had a median gestational age of 39 weeks (range 23–42 weeks); 35.4% were first-born. At age 4 years, the mean BMI and BMI z score of the sample showed a higher percentage to be obese than the standard population in the 2000 CDC growth curves: 12.4% had BMIs in the obese range and 16.5% were in the overweight range.
The univariable associations of maternal factors with child’s BMI z score at 4 years of age are shown in Table 2. GDM was associated with a significantly higher BMI z score in the offspring. Net gestational weight gain was inversely associated with child’s BMI z score, but total gestational weight gain was not. Gestational weight gain greater than recommended by the 2009 Institute of Medicine guidelines was associated with a higher BMI z score when compared with children of women who gained within the Institute of Medicine recommendations (P<.001, data not shown). Gestational hypertension, smoking during pregnancy, and gestational age at delivery, were not related to BMI z score. Several prepregnancy maternal characteristics were associated with children’s BMI z score, including maternal prepregnancy BMI, which showed a stepwise increase in BMI z score with each increase in the BMI category and a protective effect of being underweight. Chronic hypertension was associated with higher BMI z score as was Hispanic ethnicity (P<.001). Being married (P<.05) and privately insured had a negative relation with the child’s BMI.
Multivariable regression estimates for the entire sample, using adjusted net gestational weight gain as a continuous variable, are shown in Table 3. In the model, which included only prenatal factors, GDM was significantly associated with higher BMI z score of the children. When maternal prepregnancy factors were added, the association of GDM was no longer significant; there was a significant positive relation of adjusted net gestational weight gain with child’s BMI z score. Maternal prepregnancy BMI also was significantly and positively related to BMI z score. The addition of maternal demographic characteristics did not change the results for adjusted net gestational weight gain or prepregnancy BMI. Children of African American mothers and married mothers had significantly lower BMI z scores than children of white or unmarried mothers, whereas children of mothers of Hispanic ethnicity had significantly higher scores.
Gestational weight gain was also examined using the prepregnancy BMI-based categories from the 2009 Institute of Medicine recommendations for full term births (n=2,958). The mean BMI z score for children whose mothers had inadequate gestational weight gain was 0.32, whereas it was 0.40 for children of mothers with adequate gestational weight gain. Both values are significantly lower than the mean of 0.55 for children of mothers with excess gestational weight gain (P<.001). Results of the multivariable regression models for these comparisons are shown in Table 4. When only prenatal factors were included, the association of children’s BMI z scores with excess gestational weight gain was significant, but it was no longer significant in later models. After adjustment for maternal prepregnancy factors in Models 2 and 3, only the protective effect of inadequate gestational weight gain was significant.
We examined the relation of mothers’ prenatal factors with the BMI of their offspring at 4 years of age using data from linked medical records of mothers and their children. We focused on the relation of gestational weight gain, GDM, gestational hypertension, and tobacco use with children’s BMI z score while accounting for the influences of mothers’ prepregnancy health. Preconception factors had a greater influence on a child’s BMI z score than prenatal factors.
As suggested in prior research, the mother’s prepregnancy BMI confounded the association of GDM with the child’s BMI z score; no significant relation was observed between GDM and the child’s BMI after adjustment for the mother’s BMI.17 This finding is consistent with the absence of an observed protective effect of tight control of hyperglycemia during pregnancy.18
Children of mothers with inadequate gestational weight gain had a significantly lower BMI z score on average than those whose mothers gained within recommended or gained excess weight during pregnancy. The relative effect of gestational weight gain category on BMI z score of children is similar to the findings observed for 3 year olds in Project Viva using the 1990 Institute of Medicine gestational weight gain categories. The investigators found children of women who had inadequate gestational weight gain had significantly lower BMI z score than those whose mother gained an adequate or excessive amount of weight.19
Offspring of mothers of Hispanic ethnicity were more likely to be obese, consistent with national surveillance data for children.6 The lack of an association of higher average BMI z score for children whose mothers were African American is similar to findings from the Fragile Families and Child Wellbeing Survey.4 Unmeasured differences among the racial and ethnic groups in both studies may play an important role in the genesis of health disparities related to obesity in children. Being married at delivery was protective, potentially reflecting both prenatal and postnatal cultural, educational, and socioeconomic factors of each family.20,21
Factors in the intrauterine environment appear to influence future risk of cardiovascular disease through epigenetic and other cellular pathways, potentially mediated in part by the placenta.22–24 However, the relevance of earlier findings related to growth restriction to contemporary populations is unclear; the effects of excess nutrition during the prenatal period are not well established. For example, it is possible that the link between GDM and future obesity in the offspring is the result of inherited susceptibility and behavioral factors, not a result of intrauterine exposure. This alternative explanation is consistent with a growth trajectory toward obesity, which has been observed only later in childhood among the offspring of mothers with GDM.25 The child’s tendency to become obese may be the result of factors from the postnatal period acting on a susceptible host.
There are several limitations to this study. The data are from electronic medical records collected at the point of care and may include both systematic and random errors. The maternal measures of height and weight were self-reported. Prior research suggests this may be an underestimation of gestational weight gain which is likely to lead to an overestimation of the relation between gestational weight gain and BMI z score.26 The absence of information about postnatal factors during the first year such as breastfeeding or sleep may represent unmeasured confounders.27 Additionally, because the sample is not population-based the findings may not be generalizable to the United States overall. Important strengths, nevertheless, include a large and diverse sample and the use of medical records to obtain more reliable clinical data about the mother and her pregnancy and to enable exclusion of children with chronic illnesses that may affect growth.
Our results suggest risk factors present before pregnancy and during the preconception period make a greater contribution to the child's subsequent BMI, and hence risk for obesity, than those specific to the prenatal period. A significant portion of the variation in BMI z score, nevertheless, remained unexplained. The study did not address the underlying genetic contribution to BMI or the potentially strong influence of the postnatal environment suggested by twin and sibling studies.28–30
The study findings have important implications for research and clinical care. They provide further evidence to support a possible causal link between maternal prepregnancy obesity and childhood obesity. They do not help, however, discern whether this link is related to underlying genetic predisposition, maternal behaviors, or the influence of the postnatal environment.10,31–33 Our results, nevertheless, suggest that interventions focusing on the prenatal period may not effectively reduce the risk of obesity during childhood. An overall strategy to optimize women’s health before childbearing may not only be effective in reducing unfavorable pregnancy outcomes, but also improve children’s health in general.
1. Speiser PW, Rudolf MC, Anhalt H, Camacho-Hubner C, Chiarelli F, Eliakim A, et al.. Childhood obesity. J Clin Endocrinol Metab 2005;90:1871–87.
2. Withrow D, Alter DA. The economic burden of obesity worldwide: a systematic review of the direct costs of obesity. Obes Rev 2011;12:131–41.
3. Wang YC, Gortmaker SL, Taveras EM. Trends and racial/ethnic disparities in severe obesity among US children and adolescents, 1976–2006. Int J Pediatr Obes 2010 Mar 17 [Epub ahead of print].
4. Kimbro RT, Brooks-Gunn J, McLanahan S. Racial and ethnic differentials in overweight and obesity among 3-year-old children. Am J Public Health 2007;97:298–305.
5. Orsi CM, Hale DE, Lynch JL. Pediatric obesity epidemiology. Curr Opin Endocrinol Diabetes Obes 2011;18:14–22.
6. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of obesity and trends in body mass index among US children and adolescents, 1999–2010. JAMA 2012;307:483–90.
7. Nader PR, O'Brien M, Houts R, Bradley R, Belsky J, Crosnoe R, et al.. Identifying risk for obesity in early childhood. Pediatrics 2006;118:e594–601.
8. Guo SS, Wu W, Chumlea WC, Roche AF. Predicting overweight and obesity in adulthood from body mass index values in childhood and adolescence. Am J Clin Nutr 2002;76:653–8.
9. Godfrey KM, Gluckman PD, Hanson MA. Developmental origins of metabolic disease: life course and intergenerational perspectives. Trends Endocrinol Metab 2010;21:199–205.
10. Oken E. Maternal and child obesity: the causal link. Obstet Gynecol Clin North Am 2009;36:361–77, ix–x.
11. Ehrenthal DB, Jurkovitz C, Hoffman M, Jiang X, Weintraub WS. Prepregnancy body mass index as an independent risk factor for pregnancy-induced hypertension. J Womens Health (Larchmt) 2011;20:67–72.
12. Ehrenthal DB, Jurkovitz C, Hoffman M, Kroelinger C, Weintraub W. A population study of the contribution of medical comorbidity to the risk of prematurity in blacks. Am J Obstet Gynecol 2007;197:409.e1–6.
13. Clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults–the evidence report. National Institutes of Health. Obes Res 1998;6:51S–209S.
14. Rasmussen KM, Abrams B, Bodnar LM, Butte NF, Catalano PM, Maria Siega-Riz A. Recommendations for weight gain during pregnancy in the context of the obesity epidemic. Obstet Gynecol 2010;116:1191–5.
15. Smith N, Coleman KJ, Lawrence JM, Quinn VP, Getahun D, Reynolds K, et al.. Body weight and height data in electronic medical records of children. Int J Pediatr Obes 2010;5:237–42.
16. Grummer-Strawn LM, Reinold C, Krebs NF. Use of World Health Organization and CDC growth charts for children aged 0–59 months in the United States. MMWR Recomm Rep 2010;59:1–15.
17. Philipps LH, Santhakumaran S, Gale C, Prior E, Logan KM, Hyde MJ, et al.. The diabetic pregnancy and offspring BMI in childhood: a systematic review and meta-analysis. Diabetologia 2011;54:1957–66.
18. Gillman MW, Oakey H, Baghurst PA, Volkmer RE, Robinson JS, Crowther CA. Effect of treatment of gestational diabetes mellitus on obesity in the next generation. Diabetes Care 2010;33:964–8.
19. Oken E, Taveras EM, Kleinman KP, Rich-Edwards JW, Gillman MW. Gestational weight gain and child adiposity at age 3 years. Am J Obstet Gynecol 2007;196:322.e1–8.
20. Singh GK, Kogan MD, Van Dyck PC, Siahpush M. Racial/ethnic, socioeconomic, and behavioral determinants of childhood and adolescent obesity in the United States: analyzing independent and joint associations. Ann Epidemiol 2008;18:682–95.
21. Casey PH, Simpson PM, Gossett JM, Bogle ML, Champagne CM, Connell C, et al.. The association of child and household food insecurity with childhood overweight status. Pediatrics 2006;118:e1406–13.
22. Simeoni U, Barker DJ. Offspring of diabetic pregnancy: long-term outcomes. Semin Fetal Neonatal Med 2009;14:119–24.
23. Hales CN, Ozanne SE. For debate: fetal and early postnatal growth restriction lead to diabetes, the metabolic syndrome and renal failure. Diabetologia 2003;46:1013–9.
24. Ouyang F, Parker M, Cerda S, Pearson C, Fu L, Gillman MW, et al.. Placental weight mediates the effects of prenatal factors on fetal growth: the extent differs by preterm status. Obesity (Silver Spring) 2012 Apr 18 [Epub ahead of print].
25. Crume TL, Ogden L, Daniels S, Hamman RF, Norris JM, Dabelea D. The impact of in utero exposure to diabetes on childhood body mass index growth trajectories: the EPOCH study. J Pediatr 2011;158:941–6.
26. Bodnar LM, Siega-Riz AM, Simhan HN, Diesel JC, Abrams B. The impact of exposure misclassification on associations between prepregnancy BMI and adverse pregnancy outcomes. Obesity (Silver Spring) 2010;18:2184–90.
27. Gillman MW, Rifas-Shiman SL, Kleinman K, Oken E, Rich-Edwards JW, Taveras EM. Developmental origins of childhood overweight: potential public health impact. Obesity (Silver Spring) 2008;16:1651–6.
28. Branum AM, Parker JD, Keim SA, Schempf AH. Prepregnancy body mass index and gestational weight gain in relation to child body mass index among siblings. Am J Epidemiol 2011;174:1159–65.
29. The NS, Adair LS, Gordon-Larsen P. A study of the birth weight-obesity relation using a longitudinal cohort and sibling and twin pairs. Am J Epidemiol 2010;172:549–57.
30. Skidmore PM, Cassidy A, Swaminathan R, Richards JB, Mangino M, Spector TD, et al.. An obesogenic postnatal environment is more important than the fetal environment for the development of adult adiposity: a study of female twins. Am J Clin Nutr 2009;90:401–6.
31. Catalano PM, Farrell K, Thomas A, Huston-Presley L, Mencin P, de Mouzon SH, et al.. Perinatal risk factors for childhood obesity and metabolic dysregulation. Am J Clin Nutr 2009;90:1303–13.
32. Reynolds RM, Osmond C, Phillips DI, Godfrey KM. Maternal BMI, parity, and pregnancy weight gain: influences on offspring adiposity in young adulthood. J Clin Endocrinol Metab 2010;95:5365–9.
33. Pirkola J, Pouta A, Bloigu A, Hartikainen AL, Laitinen J, Jarvelin MR, et al.. Risks of overweight and abdominal obesity at age 16 years associated with prenatal exposures to maternal prepregnancy overweight and gestational diabetes mellitus. Diabetes Care 2010;33:1115–21.