Pham, Michelle T. MD; Brubaker, Katherine MD; Pruett, Kimberly MD; Caughey, Aaron B. MD, PhD
The increasing prevalence of childhood obesity has been a major topic of societal concern. To develop preventive strategies, it is important to identify which factors are associated with childhood obesity. There is increasing evidence supporting the theory that intrauterine exposure to maternal diabetes or hyperglycemia places offspring at increased risk for long-term adverse events, including obesity, diabetes, and metabolic syndrome.1 There is also evidence to show that maternal obesity, excessive gestational weight gain, and large-for-gestational-age (LGA) newborns are associated with increased risk of childhood obesity.1–5 These conditions also tend to be associated with gestational diabetes mellitus (GDM). Thus, the causal relationship between these factors and childhood obesity is uncertain and unclear.
A substantial portion of the evidence supporting the association between offspring obesity and intrauterine exposure to diabetes comes from studies examining pregestational diabetes.6,7 Studies exclusively examining GDM have been more equivocal. A recent systematic review by the Centers for Disease Control and Prevention demonstrated inconsistent evidence of an association between GDM and offspring obesity attributable to methodologic limitations of existing studies.8 These design limitations have included combining pregestational and gestational diabetes into one exposure group, lack of a nondiabetic control group, and failure to control for important potential confounders.
Given that early identification of childhood obesity could lead to earlier interventions, the primary aim of our study was to estimate the risk of obesity in the toddler offspring (aged 2–4 years) of mothers who had GDM. A secondary aim was to estimate whether childhood overweight and obesity were associated with maternal prepregnancy body mass index (BMI, calculated as weight (kg)/[height (m)]2), gestational weight gain, LGA, or race or ethnicity of mothers who had GDM.
MATERIALS AND METHODS
This was a retrospective cohort study at Kaiser Permanente Northern California in Santa Clara. This study was approved with waiver of consent by the Kaiser Foundation Research Institute Institutional Review Board in accordance with regulatory criteria. Kaiser Permanente Northern California is an integrated health care delivery system that has approximately 3.3 million members from diverse racial, ethnic, and socioeconomic backgrounds. Data were collected at the Kaiser Permanente Northern California Santa Clara medical center, which is a tertiary care teaching hospital with outpatient clinics.
The Quantitative Sentinel computerized perinatal database contained all prenatal and neonatal information for all patients receiving prenatal care at Kaiser Permanente Northern California Santa Clara. All patients had an initial examination (mean gestational age 8.3 weeks) during which data including weight, height, and demographics information were collected. The height and weight during this visit were used to calculate the BMI, which was used as the early pregnancy BMI and we also considered this an approximation to the prepregnancy BMI. Gestational weight gain was calculated from the last recorded value after 37 weeks of gestation. Race or ethnicity was categorized as white, black, Latina, Asian (Chinese, Vietnamese, Japanese, Thai, Korean, Pacific Islander, other Asian), and South Asian (Indian, Pakistani, Afghani, Bangladeshi, other South Asian). The Quantitative Sentinel database also provided neonatal outcomes, including birth weight, sex, mode of delivery, and fetal anomalies. Any demographic data not found in this database were later collected by medical record review through the Kaiser Permanente Northern California electronic medical record.
The Regional Perinatal Nursing Services database identified all patients with gestational diabetes at Kaiser Permanente Northern California Santa Clara. Kaiser Permanente Northern California universally screens for GDM using a 50-g, 1-hour glucose challenge test between 24 and 28 weeks of gestation. Women identified as high risk (personal history of pregestational or gestational diabetes, family history of diabetes, history of macrosomia) were screened during the first trimester. Those who had an elevated 1-hour glucose challenge test of 140 mg/dL or more were then evaluated with the 100-g 3-hour glucose tolerance test. The diagnosis of GDM was based on having two or more elevated values out of the four possible time points measured. Of note, Kaiser Permanente Northern California used the National Diabetes Data Group criteria for diagnosis of GDM until April 2007, when Kaiser Permanente Northern California lowered the threshold of diagnosis and switched to the Carpenter and Coustan criteria.9 At Kaiser Permanente Northern California Santa Clara, all patients with GDM participated in an educational class focusing on risks of GDM, blood glucose monitoring, dietary modifications, postprandial exercise, and appropriate weight gain during pregnancy. They also were seen at a consultation visit in the perinatal health clinic. The perinatal nursing service contacted these women weekly using telephone calls to review their blood sugar diaries. If values exceeded normal limits on more than three occasions during 1 week without obvious deviations from the recommended diet and exercise regime, or if high values persisted despite efforts to improve diet, medical treatment with glyburide or insulin was started. All patients with GDM were offered induction of labor by 40 weeks of gestation.
Maternal and neonatal data were collected for the comparison group (non-GDM patients) who delivered between 37 and 41 completed weeks of gestation from August 2004 to December 2006. Sample size was calculated to detect (with 80% power and P<.05) toddler obesity rates of 20.8% in non-GDM mother and child pairs compared with 28.4% in GDM woman and toddler pairs, with 275 pairs in the GDM group and 2,180 in the non-GDM group. To achieve the adequate sample size, data were collected for an additional year through December of 2007 for the GDM group. Exclusions for both groups known to affect neonatal birth weight were conditions including hypertensive disorders, smoking, alcohol, multiple gestation, congenital anomalies, and hyperemesis. In addition, women with more than 2+ edema were excluded because of the potential confounding effect on gestational weight gain. Although the weight gain during pregnancy recommendations by the Institute of Medicine were unclear regarding women with GDM, these guidelines were used for both study groups (GDM and non-GDM) as follows: BMI 18.5–24.9, gain 25–35 pounds; BMI 25–29.9, gain 15–25 pounds; and BMI 30 or more, gain 11–20 pounds.10 Adequacy of weight gain was defined in relation to the Institute of Medicine guideline for each BMI category, with inadequate gain being less than their recommendations, and with excessive gain being more than the Institute of Medicine guideline. Neonatal LGA was defined as newborn birth weight more than the 90th percentile based on Kaiser Permanente Northern California--specific growth charts. Data for the toddler offspring were gathered by chart review of the electronic medical records approximately 2–4 years after birth. Height and weight were routinely collected at well-child visits. The child's BMI percentile for gender and age was calculated based on the Centers for Disease Control and Prevention growth chart formulas.11 Overweight and obesity were defined as BMI more than the 85th percentile. If height was not recorded, weight more than the 90th percentile was used to define overweight and obesity.
Statistical analysis was performed using the SAS Statistical Analysis System for Windows 9.13, and P<.05 was considered significant. Differences by GDM status were analyzed with Pearson χ2 test for categorical variables and Student t test for continuous variables. Differences in rates of GDM, LGA, and toddler overweight or obesity by race or ethnicity were examined by χ2 test adjusted for multiple comparisons using Hochberg method.12 Multivariable logistic regression models to predict toddler overweight and obesity were used to identify factors predictive of toddler overweight and obesity with maternal prepregnancy BMI as the main risk factor. Potential confounders included in all models were maternal age, race, height, and child age. The base model comprised the main risk factor and the confounders followed by three additional models, and each included one or more risk factors contained in the model, specifically, maternal GDM, maternal GDM and LGA, and maternal GDM and maternal weight gain by BMI category.
A total of 2,644 women met the study cohort selection criteria, including 459 (17.4%) women in the GDM group and 2,185 (82.6%) in the non-GDM group. Women with GDM had a higher mean BMI of 27.4 (standard deviation 6.1) than did those without GDM (mean 24.6, standard deviation 4.7; P<.001; Table 1). However, women with GDM gained less weight in pregnancy than those without GDM (20.2 pounds compared with 29.6 pounds; P<.001; Table 1). Among all maternal prepregnancy BMI groups, fewer women in the GDM group gained excess weight as defined by the Institute of Medicine guidelines (Table 2). Mean birth weights of newborns of women with GDM was not different compared with those of newborns of women without GDM (3,406 g compared with 3,404 g; P=.93; Table 1). Women with GDM also delivered slightly earlier than did those without GDM (39.3 weeks compared with 39.6 weeks; P<.001; Table 1).
Of the 459 women with GDM, we were able to obtain toddler weight for 255 (55.6%) offspring, hereafter termed “woman and toddler pairs” of this group, whereas such data were collected from 84.1% (1,838 of 2,185) of the non-GDM group. Toddler weight data for the remaining woman and toddler pairs were unavailable primarily because of loss to follow-up. When comparing the BMI percentiles of the toddlers of the women with GDM with those without GDM (Table 3), the slight difference was not statistically different (51.8 compared with 55.2 percentile; P=.12; Table 3) and the rates of overweight and obesity between the two groups also did not differ (23.9% compared with 23.5%; P=.88; Table 3).
Of interest, the rate of GDM was higher in Latina and both Asian subgroups as compared with white women. However, the rate of LGA in newborns was lower in South Asian women as compared with white women. Similarly, the rate of toddler obesity was lower in South Asians than in whites (Fig. 1).
Rates of gestational...Image Tools
Given that maternal obesity, excessive gestational weight gain, and LGA each have been separately associated with childhood obesity, we performed three multivariable models examining their effects on childhood obesity in relation to GDM. The adjusted logistic regression models controlled for maternal age, height, race or ethnicity, and child age. The base multivariable logistic regression model (model A) demonstrated that, when controlling for the potential confounders (maternal age, height, race or ethnicity, and child age), toddlers of women with higher maternal prepregnancy BMIs had higher odds of childhood overweight and obesity. In all four models, the odds of toddler overweight and obesity in the maternal prepregnancy BMI 30 or more group remained, approximately, at 2.5 times the odds compared with normal-weight mothers (Table 4). Gestational diabetes mellitus was not associated with increased risk of toddler overweight and obesity (model B in Table 4). Although LGA also increased the odds of toddler obesity (adjusted odds ratio [OR] 1.8, 95% confidence interval 1.4–2.3; model C in Table 4), excessive gestational weight gain did not (adjusted odds ratio 1.1, 95% confidence interval 0.8–1.4; model D in Table 4).
This study examined the risk of childhood obesity in the toddler offspring of women with GDM compared with that in the toddler offspring of women without GDM. We found that, among our diverse cohort of 255 women with GDM and their offspring, there was no difference in the rate of overweight and obesity at 2–4 years of age compared with the toddlers of women without GDM.
In our results we found that childhood obesity at 2–4 years of age was not associated with GDM but was associated with maternal BMI and neonatal LGA, as expected. Although previous studies of older offspring of women with GDM have shown some association with childhood obesity,1,2 in this age group we did not see a difference. One previous study of 51 children at 3 years of age also did not show a difference in BMI but did show an increase in adiposity.5 We suspect that one potential reason for the lack of difference in childhood obesity rates between the GDM and non-GDM groups was attributable to the well-controlled nature of GDM in the women in our cohort. All women in Kaiser Permanente Northern California have insurance that allows for full prenatal coverage. Kaiser Permanente Northern California has a strict regimented program for screening, follow-up, and treatment of GDM. The education and close follow-up provided by our Regional Perinatal Nursing Services likely explains how the women with GDM gained significantly less weight than did those without GDM. This was most notable in the overweight (BMI 24.9–29.9) and obese (BMI 30 or higher) women, who were less than half as likely to gain excess weight compared with the women without GDM (per Institute of Medicine guidelines). This is important because excessive maternal weight gain in conjunction with gestational hyperglycemia leads to macrosomic neonates.13 In contrast, we did not find a difference in birth weight between the offspring of women with and without GDM. Thus, the significant improvement in weight gain in our study may be a major contributor to why we did not see increased birth weight or an increase in the risk of childhood obesity at ages 2–4 years.
From the Hyperglycemia and Adverse Pregnancy Outcomes study, we know that GDM increases the risk of LGA and macrosomic neonates.14 The Maternal-Fetal Medicine Units Network treatment of mild GDM clinical trials subsequently showed that diagnosis and treatment of GDM improves pregnancy outcomes, including reduction in LGA.15 This is consistent with our finding that there was no difference in LGA rates between our women with GDM and women without GDM. Given that appropriate GDM management can reduce rates of LGA, there is the possibility that appropriate GDM management also may affect childhood obesity. One other potential etiology of the lack of effect is that it is also possible that the intensive monitoring of GDM in pregnancy had a long-term effect on behavioral or dietary modifications for women and their offspring that offset the increased baseline risk for obesity.
This study also specifically studied children 2–4 years of age, which is unique compared with other studies in that it examined children before adiposity rebound. The adiposity rebound corresponds to the second increase in the pediatric BMI curve that occurs between ages 5 and 7 years. The adiposity rebound already has been identified as an indicator predicting obesity.4 Other studies already have suggested that children begin gaining excess weight after age 5 years.16 Therefore, it was important for us to specifically examine ages 2–4 years before the adiposity rebound. At this age, environmental factors are less likely to contribute to the association between childhood obesity and in utero factors, such as gestational diabetes. Longitudinal follow-up of this same cohort during the ages of adiposity rebound could result in different findings.
This study has important strengths. This is one of the largest studies examining the risk of childhood obesity exclusively in GDM offspring younger than age 5 years. Most importantly, our study controlled for potential confounders, including maternal BMI, weight gain, and LGA, which enabled us to show that childhood obesity is likely more associated with maternal BMI and LGA than well-controlled GDM. Despite its strengths, this study also has some limitations. We were unable to obtain the values of glycemic control, which would allow us to stratify patients by level of hyperglycemia. Furthermore, during the study period the values of the glucose tolerance test that were used to diagnose GDM were changed from the National Diabetes Data Group criteria to the Carpenter and Coustan criteria. However, given our negative findings, it seems unlikely that either categorization would have been associated with childhood obesity on its own. In addition, our study population is only representative of our facility; thus, there was an overrepresentation of Asian and South Asian populations and an underrepresentation of the black population. Given that all individuals in our patient population are insured, this study also may not be generalizable to populations with less access to early and comprehensive prenatal care. In addition, we do not know the characteristics of the woman and toddler pairs who changed insurance and were not included in the mother–toddler calculations. As with all observational studies, our study was a retrospective cohort study that may be prone to confounding bias. Although we used statistical techniques to control for important confounders such as maternal age, BMI, and race or ethnicity, there may be residual confounders that we could not observe or for which we did not control.
In summary, among a diverse insured patient population, childhood obesity in children aged 2–4 years was not associated with GDM. These data do confirm that toddler obesity is associated with maternal BMI and LGA, which are associated with poorly controlled GDM. This may suggest that when controlling for maternal BMI, it is not the actual attribute of having GDM that contributes to childhood obesity, but possibly that GDM management may be a modifiable risk factor for childhood obesity. More research is needed to determine whether the interventions of GDM programs or those targeted at decreasing prepregnancy maternal BMI could help reduce the childhood obesity epidemic.
1. Boney CM, Verma A, Tucker R, Vohr BR. Metabolic syndrome in childhood: association with birth weight, maternal obesity, and gestational diabetes mellitus. Pediatrics 2005;115:e290–6.
2. Gillman MW, Rifas-Shiman S, Berkey CS, Field AE, Colditz GA. Maternal gestational diabetes, birth weight, and adolescent obesity. Pediatrics 2003;111:221–6.
3. Von Kries R, Ensenauer R, Beyerlein A, Amann-Gassner U, Hauner H, Rosario AS. Gestational weight gain and overweight in children: results from the cross sectional German KiGGS study. Int J Pediatr Obes 2010;6:45–52.
4. Rolland-Cachera MF, Deheeger M, Bellisle F, Sempe M, Guilloud-Bataille M, Patois E. Adiposity rebound in children: a simple indicator for predicting obesity. Am J Clin Nutr 1984;39:129–35.
5. Wright CS, Rifas-Shiman SL, Rich-Edwards JW, Taveras EM, Gillman MW, Oken E. Intrauterine exposure to gestational diabetes, child adiposity, and blood pressure. Am J Hypertens 2009;22:215–20.
6. Pettit DJ, Bennett PH, Knowler WC, Baird HR, Aleck KA. Gestational diabetes mellitus and impaired glucose tolerance during pregnancy. Long-term effects on obesity and glucose tolerance in the offspring. Diabetes 1985;34:119–22.
7. Silverman BL, Rizzo T, Green OC, Cho NH, Winter RJ, Ogata ES, et al.. Long-term prospective evaluation of offspring of diabetic mothers. Diabetes 1991;40:121–5.
8. Kim SY, England JL, Sharma A, Njoroge T. Gestational diabetes mellitus and risk of childhood overweight and obesity in offspring: a systematic review. Exp Diabete Res 2011;2011:541308.
9. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 2000;23:S4–19.
10. Rasmussen KM, Yaktine AL, editors. Weight gain during pregnancy: reexamining the recommendations. Washington, DC: The National Academies Press; 2009.
11. Kuczmarski RJ, Ogden CL, Guo SS, Grummer-Strawn LM, Flegal KM, Mei Z, et al.. 2000 CDC growth charts for the United States: Methods and development. National Center for Health Statistics. Vital Health Stat 2002;246:1–190.
12. Hochberg Y, Benjamini Y. More powerful procedures for multiple significance testing. Stat Med 1990;9:811–8.
13. Hillier TA, Pedula KL, Vesco KK, Schmidt MM, Mullen JA, LeBlanc ES, et al.. Excess gestational weight gain: modifying fetal macrosomia risk associated with maternal glucose. Obstet Gynecol 2008;112:1007–14.
14. Metzger BE, Lowe LP, Dyer AR, Trimble ER, Chaovarindr U, et al.. Hyperglycemia and adverse pregnancy outcomes. N Engl J Med 2008;358:1991–02.
15. Landon MB, Spong CY, Thom E, Carpenter MW, Ramin SM, Casey B, et al.. A multicenter, randomized trial of treatment of mild gestational diabetes. N Engl J Med 2009;361:1339–48.
16. Williams SM, Goulding A. Patterns of growth associated with the timing of adiposity rebound. Obesity (Silver Spring) 2008;17:335–41.
Figure. No available...Image Tools