Abstract
Objective
An association between in utero exposure to gestational glucose intolerance ([GGI], abnormal glucose screening without gestational diabetes), and offspring obesity has not been consistently observed.
Methods
In a retrospective cohort, we studied the risk of obesity (body mass index [BMI] ≥ 95th percentile), in 2–5, 6–10, and 11–18-year-olds exposed to varying degrees of maternal glycemia in utero: normal glucose tolerance (NGT), GGI (0 abnormal glucose values, GGI-0, or 1 abnormal value, GGI-1) or gestational diabetes (GDM ≥2 of 4 abnormal values). We used generalized estimating equations for logistic regression to estimate odds ratios for obesity in each glycemic category compared to NGT, adjusting for maternal age, parity, insurance, race/ethnicity, marital status, infant sex, gestational age, and gestational weight gain. A second model additionally adjusted for maternal 1st trimester BMI.
Results
We included 27,876 children and adolescents from 23,334 (83.7%) NGT pregnancies, 3,413 (12.2%) GGI pregnancies, and 1,129 (4.1%) GDM pregnancies. The prevalence of obesity was 13.5% at age 2–5, 20.3% at age 6–10, and 23.4% at age 11–18. Those exposed to GGI-1 and GDM had increased odds of obesity compared to NGT. Adjusting for maternal BMI attenuated this association in all age and glycemic exposure groups, but it remained significantly elevated in 6–10-year-olds exposed to GDM (odds ratio (OR): 1.21, 95%CI [1.01, 1.46] and 11–18-year-olds exposed to GGI-1 and GDM (GGI-1 OR: 1.44 [1.14, 1.81]; GDM OR: 1.28 [1.03, 1.59]).
Conclusion
Older children and adolescents exposed to GGI-1 and GDM in utero have a higher risk of obesity than those born to NGT pregnancies, even after accounting for maternal BMI.
Keywords: gestational diabetes, pediatric obesity, subclinical hyperglycemia
Introduction
The obesity epidemic affects children of all age ranges. The prevalence of obesity (defined as a body mass index [BMI] ≥95th percentile for age and sex) increases with age. In the United States current prevalence is close to 20%, ranging from 13% in 2–5 year-olds to 22% in adolescence.1 Children with obesity have a higher risk of adult obesity2 and cardiometabolic consequences; as a result of the obesity epidemic, these consequences are being diagnosed earlier. Early onset of cardiometabolic disease is associated with worse outcomes.3 Identifying modifiable risk factors is essential to prevention. Pregnancy is a critical period in development and fetal exposures can impact lifelong risk of disease.4
Hyperglycemia in pregnancy, including levels below the threshold for gestational diabetes (GDM), is associated with perinatal adverse outcomes.5,6 Pre-existing diabetes and GDM is associated with long term risk of offspring obesity.7 In early childhood higher risk of obesity in offspring born to GDM pregnancies is attenuated when adjusting for maternal BMI8, so it is not clear whether GDM is merely a risk marker for an obesogenic environment or if exposure to glycemia in utero itself increases risk. A study in 236 children of individuals with untreated GDM and 480 children born from a pregnancy with gestational glucose intolerance (abnormal screening with normal 3-hour diagnostic oral glucose tolerance test) found an association between oral glucose tolerance test (OGTT) levels and adiposity (measured by sum of skinfolds and subscapular/triceps ratio) but not with BMI in childhood (5–10 years).9 However, BMI is a widely used tool in clinical practice. On the other hand, an observational study looking exclusively at late adolescence (16–20 years old) demonstrated increased odds of overweight and obesity in individuals with gestational glucose intolerance (GGI) even after adjusting for maternal BMI in adolescence.10 Treatment of GDM has been associated with decreased risk of adverse outcomes during the perinatal period, but the protective effect on long term risk is less clear.6,11 Understanding the association between treated and untreated mild hyperglycemia in pregnancy and the risk of overweight/obesity in different stages of childhood and adolescence could inform interventions in pregnancy to prevent childhood and adult disease.
In this study, we investigated the risk of overweight and obesity in children across different age groups with exposure to different degrees of hyperglycemia in pregnancy to describe the effects across childhood and adolescence.
Methods
The Massachusetts General Hospital (MGH) Maternal Health Cohort (MHC) includes delivery information on 56,107 pregnancies derived from electronic health record data. In this retrospective cohort study, we linked prenatal data with anthropometric measurements of offspring followed within Massachusetts General Brigham (MGB), extracted from the MGB Research Patient Data Registry. The MGB Institutional Review Board approved the use of electronic health records for this study and waived the requirement for informed consent.
We included pregnancies that could be linked to children and adolescents with weight and height measurements taken between 2 and 18 years of age. We excluded pregnancies with incomplete GDM screening and multiple gestation pregnancies. We excluded children delivered prior to 28 weeks’ gestation. Individuals with known pre-existing diabetes do not undergo screening for GDM, and therefore, they were not included in this study12.
We removed implausible anthropometric records, which were those meeting at least one of the following criteria: 1) weight and height measurements more than four standard deviations from the age- and sex-specific population mean, 2) BMI percentiles that differed by more than 50 from the previous and following percentile in the same individual, 3) height measurements that were two inches larger or smaller from both the previous and the following measurement in the same individual.
Screening for GDM at the institution from which the study population was derived used a two-step approach, congruent with American College of Obstetrics and Gynecology guidelines13: screening with a 1-hour 50 g glucose loading test (GLT), and if blood glucose was ≥140 mg/dl, diagnostic testing with a 3-hour 100g OGTT. For this study, we applied the Carpenter-Coustan criteria to define GDM, which is currently used in most clinical settings in the United States.14 Clinically, during the period of study, GDM diagnosis and treatment was determined using National Diabetes Data Group (NDDG) criteria, which have higher thresholds for diagnosis.15 Normal glucose tolerance was defined as screening test (GLT) <140 mg/dl. We considered individuals with 0–1 abnormal values on OGTT (after GLT 1h-glucose >140mg/dL) to have GGI. We divided GGI into subtypes: GGI with an elevated GLT 1h-glucose and a normal OGTT and GGI with an elevated GLT 1h-glucose and a 1 abnormal OGTT value (Figure 1). Individuals with at least 2 abnormal values on diagnostic OGTT were classified as GDM13. This group was further divided into those who received a clinical diagnosis using NDDG criteria and thus were treated (NDDG GDM) and those who met criteria for GDM using Carpenter-Coustan criteria for the purpose of the current study but did not receive treatment (CC GDM).
Figure 1. Gestational Glucose Intolerance Classification and Exposure groups.
Abbreviations: GLT: glucose loading test, OGTT: oral glucose tolerance test, GDM: gestational diabetes.
Offspring outcomes:
Overweight/obesity was defined as a BMI ≥85th percentile for age and sex, and obesity as a subset of the aforementioned overweight/obesity outcome with a BMI ≥95th percentile for age and sex. BMI percentiles were based on The Centers for Disease Control and Prevention 2000 growth charts and pediatric guidelines.16,17 The analysis was conducted in three age ranges: 2–5, 6–10, and 11–18 years. In each age range, all measurements within the range were used, and the median BMI percentile across those measurements were used to define the outcome in each individual. Most children contributed data to multiple age groups.
Covariates:
We adjusted for potential confounders including maternal age, gestational weight gain, parity, insurance, race/ethnicity, marital status, and gestational age at delivery. Separately, we additionally adjusted for first trimester maternal BMI which was derived using standardized measurements to 12 weeks’ gestation obtained through a previously described interpolation/extrapolation procedure.6,18 This approach was also used to estimate gestational weight gain by subtracting the standardized 12-week weight from the last weight prior to delivery. Parity was binarized (nulliparous or not). Marital status was divided into married/partnered versus not married. Individuals could identify as Asian, Black, Latino, or White, multiple categories or none of the above. Insurance was defined as private, public, or limited/none.
Statistical Analyses:
We used generalized estimating equations for logistic regression to compare the odds of overweight/obesity or obesity in children from GGI pregnancies to those from pregnancies with normal glucose tolerance. We calculated the odds of overweight/obesity and obesity for each of the three age ranges (2–5, 6–10, and 11–18) and the risk in children exposed to GGI, GGI subtypes (GGI with a normal OGTT and GGI and 1 abnormal value) as well as GDM and GDM subtypes (untreated CC GDM and treated NDDG GDM) compared to normal glucose tolerance.
Generalized estimating equations were used to account for multiple pregnancies from the same individual.
We used R version 4.3.1 for analyses19 and fit generalized estimating equations using the glmgee function from the geepack package in R.20–22
Sensitivity Analyses:
We performed sensitivity analysis where we only included children/adolescents that had at least 3 available BMI measurements within an age category as defined above. This stricter inclusion criteria was used to assess the internal validity of our results. Additionally, we examined the results by offspring sex. Finally, we performed a post hoc stratified analysis by maternal BMI (<25 kg/m2 and ≥25 kg/m2).
Results
The final cohort included 27,876 singleton deliveries born after 28 weeks’ gestation with available BMI data between 2 and 18 years of age and complete GDM screening (Figure 2). A total of 23,334 (83.7%) pregnancies were characterized as normal glucose tolerance, 3,413 (12.2%) as GGI, and 1,129 (4.1%) as GDM. The GGI group was divided into 2,525 (9.0%) pregnancies with GGI with a normal 3h-OGTT and 888 (3.2%) with GGI and 1 abnormal value (Figure 1). In total, 18,644 (66.9% of pregnancy cohort) deliveries had BMI data at ages 2–5 years, 17,238 (61.8%) had data at ages 6–10 years and 12,558 (45.0%) had data from ages 11–18 years. Individuals excluded had an increased likelihood of being married/partnered, non-Hispanic White, and having private insurance.
Figure 2. Cohort used for Analysis.
Abbreviations: MGH: Mass General Hospital, BMI: body mass index, GDM: gestational diabetes
Table 1 describes baseline characteristics of the individual pregnancies by exposure category. First trimester maternal BMI was greater in the GGI with 1 abnormal value and GDM groups while gestational weight gain and gestational age at delivery were lowest in the GDM group, likely due to clinical interventions indicated during pregnancy. The GGI with 1 abnormal value and GDM groups were more likely to identify as non-Hispanic Black or as Latino, and to have public insurance. Absolute birth weight and prevalence of large for gestational age birth weight were greater in the GGI with 1 abnormal value and GDM groups, as previously reported.6
Table 1.
Characteristics of study cohort
| Normal Glucose Tolerance (NGT) | Positive GLT and 0–1 Abnormal OGTT Values (GGI) | Positive GLT and 0 Abnormal OGTT Values | Positive GLT and 1 Abnormal OGTT Values | Gestational Diabetes (GDM) | |
|---|---|---|---|---|---|
| N=23,334 | N= 3,413 | N= 2,525 | N= 888 | N= 1,129 | |
| Mean (SD) or n (%) | Mean (SD) or n (%) | Mean (SD) or n (%) | Mean (SD) or n (%) | Mean (SD) or n (%) | |
| Maternal Age (years) | 29.9 (6.1) | 31.4 (5.7) | 31.2 (5.7) | 32.1 (5.7) | 32.2 (5.5) |
| Nulliparous | 11,859 (51%) | 1,626 (48%) | 1,230 (49%) | 396 (45%) | 513 (45%) |
| First trimester BMI (kg/m2) * | 25.4 (5.2) | 26.1 (5.4) | 25.7 (5.1) | 27.3 (5.8) | 28.4 (6.2) |
| Gestational Weight Gain (lbs.) ϯ | 29.2 (9.4) | 28.3 (9.2) | 28.3 (8.8) | 28.4 (10.1) | 24.8 (9.5) |
| Married/Partnered | 15,334 (66%) | 2,355 (69%) | 1,735 (69%) | 620 (70%) | 760 (67%) |
| Race/Ethnicity ¥ | |||||
| Non-Hispanic Asian | 2,018 (9%) | 423 (12%) | 317 (13%) | 106 (12%) | 161 (14%) |
| Non-Hispanic Black | 1,948 (8%) | 209 (6%) | 134 (5%) | 75 (8%) | 113 (10%) |
| Hispanic/Latino | 3,514 (15%) | 488 (14%) | 348 (14%) | 140 (16%) | 192 (17%) |
| Non-Hispanic White | 13,135 (56%) | 1,839 (54%) | 1,399 (55%) | 440 (50%) | 501 (44%) |
| None of the above | 3,256 (14%) | 520 (15%) | 375 (15%) | 145 (16%) | 192 (17%) |
| Multiracial | 534 (2%) | 65 (2%) | 48 (2%) | 17 (2%) | 30 (3%) |
| Insurance | |||||
| Private | 13,486 (58%) | 2,001 (59%) | 1,511 (60%) | 490 (55%) | 557 (49%) |
| Public | 7,553 (32%) | 1,082 (32%) | 780 (31%) | 302 (34%) | 446 (40%) |
| Limited/None | 2,295 (10%) | 330 (10%) | 234 (9%) | 96 (11%) | 126 (11%) |
| Gestational age at delivery (weeks) | 39.4 (1.6) | 39.4 (1.6) | 39.5 (1.6) | 39.2 (1.8) | 38.9 (1.9) |
| Neonatal sex: Female | 11,322 (49%) | 1,646 (48%) | 1,197 (47%) | 449 (51%) | 528 (47%) |
| Birth weight (grams) | 3,363.2 (515.5) | 3,426.3 (526.2) | 3,412.1 (513.1) | 3,467.0 (560.0) | 3,427.1 (597.0) |
| Birth weight percentile | 48.0 (27.6) | 52.6 (27.8) | 51.1 (27.7) | 56.8 (27.6) | 56.2 (28.4) |
| LGA status ** | 1,718 (7%) | 337 (10%) | 221 (9%) | 116 (13%) | 170 (15%) |
Results for maternal age, first trimester BMI and gestational weight gain expressed in Mean (SD). The rest as N(%). Abbreviations: BMI: body mass index. LGA: large for gestational age birth weight; OGTT: Oral glucose tolerance test.
Interpolating/extrapolating back to 12 weeks for all missing data or data after 12 weeks GA
12-week extrapolated weight minus weight at the time of delivery,
Each person was able to select more than one race/ethnicity category.
Large for gestational age at delivery was defined as birth weight >90th percentile based on national growth curves.32
Missing data: There were 390 pregnancies missing first trimester BMI in NGT, 51 in GGI, and 21 in GDM groups. There were 21 pregnancies missing neonatal sex in the NGT, 2 in GGI and 2 in GDM groups. There were 25 pregnancies missing birth weight percentile in NGT, 3 IN GGI and 2 in GDM groups.
The prevalence of overweight/obesity and obesity increased across age groups: with 28% and 14% for 2–5-year-olds, 37% and 20% among 6–10-year-olds, and 41% and 23% in 11–19-year-olds, respectively (Figure 3).
Figure 3. Prevalence of overweight and obesity by pregnancy glucose tolerance category for each age group.
Abbreviations: NGT: normal glucose tolerance, GDM: gestational diabetes.
Adjusted models showed progressively higher point estimates for the odds of overweight/obesity and obesity with increasing age and severity of glucose exposure (Table 2, Model A Figure 4). In Model B, the odds ratios were attenuated after adjusting for maternal first trimester BMI (Model B Figure 4). However, the lower bound of the 95% CI exceeded 1 for these maternal BMI-adjusted odds ratios in older children and adolescents. In 6–10-year-olds, the odds ratio for obesity after adjustment for maternal BMI and other potential confounders in those exposed to a GDM pregnancy was 1.21, 95% CI [1.01, 1.46]. In 11–18-year-olds, we observed a higher risk of overweight/obesity in the GDM group (OR 1.33 [1.09, 1.63]) and of obesity in the GGI with 1 abnormal value (OR 1.44 [1.14, 1.81]) and GDM (OR 1.28 [1.03, 1.59]) groups (see Model B in Figure 4) after adjustment for maternal BMI.
Table 2.
Estimated odds ratios and 95% confidence intervals for the association between exposure to GGI, GDM and child BMI status
| GGI vs. NGT | GGI Subtypes | GDM vs. NGT | GDM Subtypes | ||||
|---|---|---|---|---|---|---|---|
| 0 Abnormal OGTT Values vs. NGT | 1 Abnormal OGTT Value vs. NGT | CC GDM vs. NGT | NDDG GDM | ||||
| Overweight/Obesity (BMI ≥85th percentile) at 2–5 years old | |||||||
| Unadjusted for maternal BMI* | Odds Ratio | 1.06 | 1.00 | 1.23 | 1.24 | 1.23 | 1.25 |
| (95% CI) | (0.96, 1.17) | (0.90, 1.12) | (1.03, 1.47) | (1.06, 1.46) | (0.96, 1.57) | (1.02, 1.54) | |
| Adjusted for maternal BMI** | Odds Ratio | 1.00 | 0.97 | 1.08 | 1.00 | 0.96 | 1.04 |
| (95% CI) | (0.90, 1.11) | (0.86, 1.09) | (0.90, 1.29) | (0.85, 1.18) | (0.74, 1.24) | (0.84, 1.28) | |
| Obesity (BMI ≥95th percentile) at 2–5 years old | |||||||
| Unadjusted for maternal BMI* | Odds Ratio | 1.10 | 1.01 | 1.35 | 1.48 | 1.56 | 1.42 |
| (95% CI) | (0.97, 1.25) | (0.87, 1.18) | (1.08, 1.69) | (1.21, 1.80) | (1.16, 2.09) | (1.10, 1.83) | |
| Adjusted for maternal BMI** | Odds Ratio | 1.03 | 0.97 | 1.16 | 1.18 | 1.21 | 1.15 |
| (95% CI) | (0.90, 1.17) | (0.83, 1.14) | (0.92, 1.46) | (0.96, 1.44) | (0.89, 1.64) | (0.89, 1.50) | |
| Overweight/Obesity (BMI ≥85th percentile) at 6–10 years old | |||||||
| Unadjusted for maternal BMI* | Odds Ratio | 1.19 | 1.10 | 1.44 | 1.44 | 1.38 | 1.48 |
| (95% CI) | (1.08, 1.30) | (0.99, 1.23) | (1.21, 1.71) | (1.24, 1.68) | (1.09, 1.76) | (1.22, 1.79) | |
| Adjusted for maternal BMI** | Odds Ratio | 1.09 | 1.06 | 1.19 | 1.09 | 1.04 | 1.13 |
| (95% CI) | (0.99, 1.21) | (0.94, 1.19) | (0.99, 1.42) | (0.93, 1.29) | (0.80, 1.35) | (0.92, 1.38) | |
| Obesity (BMI ≥95th percentile) at 6–10 years old | |||||||
| Unadjusted for maternal BMI* | Odds Ratio | 1.19 | 1.09 | 1.47 | 1.64 | 1.49 | 1.75 |
| (95% CI) | (1.06, 1.33) | (0.96, 1.25) | (1.21, 1.79) | (1.38, 1.95) | (1.13, 1.96) | (1.41, 2.17) | |
| Adjusted for maternal BMI** | Odds Ratio | 1.08 | 1.05 | 1.17 | 1.21 | 1.08 | 1.30 |
| (95% CI) | (0.96, 1.21) | (0.91, 1.20) | (0.95, 1.44) | (1.01, 1.46) | (0.81, 1.46) | (1.04, 1.63) | |
| Overweight/Obesity (BMI ≥85th percentile) at 11–18 years old | |||||||
| Unadjusted for maternal BMI* | Odds Ratio | 1.23 | 1.17 | 1.42 | 1.78 | 1.73 | 1.81 |
| (95% CI) | (1.10, 1.37) | (1.03, 1.32) | (1.17, 1.74) | (1.48, 2.13) | (1.29, 2.32) | (1.44, 2.27) | |
| Adjusted for maternal BMI** | Odds Ratio | 1.13 | 1.11 | 1.20 | 1.33 | 1.23 | 1.40 |
| (95% CI) | (1.01, 1.27) | (0.97, 1.26) | (0.97, 1.48) | (1.09, 1.63) | (0.89, 1.70) | (1.09, 1.79) | |
| Obesity (BMI ≥95th percentile) at 11–18 years old | |||||||
| Unadjusted for maternal BMI* | Odds Ratio | 1.27 | 1.14 | 1.71 | 1.75 | 1.71 | 1.78 |
| (95% CI) | (1.12, 1.44) | (0.98, 1.32) | (1.37, 2.12) | (1.43, 2.14) | (1.23, 2.37) | (1.39, 2.28) | |
| Adjusted for maternal BMI** | Odds Ratio | 1.17 | 1.08 | 1.44 | 1.28 | 1.24 | 1.30 |
| (95% CI) | (1.03, 1.34) | (0.93, 1.26) | (1.14, 1.81) | (1.03, 1.59) | (0.86, 1.78) | (1.00, 1.70) | |
Model A – Unadjusted for maternal BMI: Adjusted for maternal age, parity, insurance, race/ethnicity, marital status, infant sex, gestational age, gestational weight gain (GWG)
Model B – Adjusted for all covariates in Model A in addition to 1st trimester maternal BMI
Figure 4. Estimated odds ratios and 95% confidence intervals for the association between exposure to GGI, GDM and risk of obesity (BMI>95th percentile) in childhood and adolescence.
Model A (left): adjusted for maternal age, insurance type, race/ethnicity, marital status, gestational age at delivery, gestational weight gain, and parity; Model B (right): adjusted for maternal 1st trimester BMI in addition to covariates in Model A
The odds ratios in the treated (NDDG GDM) and untreated (CC GDM) groups were similar when compared to pregnancies with normal glucose tolerance (Table 2).
Sensitivity analysis including only offspring with 3 or more available BMI values demonstrated a similar trend despite the smaller sample size (Figure 5, Models A and B). Stratified analyses examining effects by offspring sex, did not find any evidence of differences in associations between GGI and GDM and childhood obesity in males versus females (data not shown).
Figure 5. Estimated odds ratios and 95% confidence intervals for the association between exposure to GGI, GDM and risk of obesity (BMI>95th percentile), for subsample with 3 or more BMI values during childhood and/or adolescence.
Model A (left): adjusted for maternal age, insurance type, race/ethnicity, marital status, gestational age at delivery, gestational weight gain, and parity; Model B (right): adjusted for maternal 1st trimester BMI in addition to covariates in Model A. The sensitivity analysis included 16,393 children in the 2–5-year-old category, 1449 in the 6–10-year-old category, and 591 in the 11–18-year-old category.
Post hoc analysis by maternal BMI, suggested a stronger association in those pregnancies with first trimester BMI ≥25 kg/m2 although confidence intervals almost completely overlapped (Table 3).
Table 3.
Estimated odds ratios and 95% confidence intervals for the association between exposure to GGI, GDM and child BMI status by maternal body mass index
| GGI vs. NGT | GGI Subtypes | GDM vs. NGT | GDM Subtypes | ||||
|---|---|---|---|---|---|---|---|
| 0 Abnormal OGTT Values vs. NGT | 1 Abnormal OGTT Value vs. NGT | CC GDM vs. NGT | NDDG GDM | ||||
| Overweight/Obesity (BMI >85th percentile) at 2–5 years old | |||||||
| Maternal first trimester BMI <25 kg/m2 | Odds Ratio | 1.02 | 1.03 | 0.97 | 0.85 | 0.86 | 0.85 |
| (95% CI) | (0.86, 1.20) | (0.86, 1.23) | (0.69, 1.37) | (0.60, 1.21) | (0.47, 1.56) | (0.55, 1.30) | |
| Maternal first trimester BMI ≥ 25 kg/m2 | Odds Ratio | 1.09 | 1.03 | 1.24 | 1.25 | 1.19 | 1.28 |
| (95% CI) | (0.97, 1.24) | (0.89, 1.19) | (1.01, 1.52) | (1.04, 1.49) | (0.91, 1.57) | (1.02, 1.61) | |
| Obesity (BMI >95th percentile) at 2–5 years old | |||||||
| Maternal first trimester BMI <25 kg/m2 | Odds Ratio | 0.87 | 0.85 | 0.93 | 0.90 | 0.97 | 0.86 |
| (95% CI) | (0.68, 1.12) | (0.64, 1.13) | (0.55, 1.57) | (0.53, 1.53) | (0.43, 2.17) | (0.44, 1.70) | |
| Maternal first trimester BMI ≥ 25 kg/m2 | Odds Ratio | 1.26 | 1.14 | 1.56 | 1.45 | 1.30 | 1.56 |
| (95% CI) | (1.10, 1.45) | (0.96, 1.34) | (1.25, 1.96) | (1.19, 1.76) | (0.96, 1.75) | (1.22, 2.00) | |
| Overweight/Obesity (BMI >85th percentile) at 6–10 years old | |||||||
| Maternal first trimester BMI <25 kg/m2 | Odds Ratio | 1.15 | 1.04 | 1.63 | 0.99 | 1.01 | 0.96 |
| (95% CI) | (0.75, 1.79) | (0.64, 1.71) | (0.71, 3.76) | (0.38, 2.59) | (0.26, 3.95) | (0.25, 3.68) | |
| Maternal first trimester BMI ≥ 25 kg/m2 | Odds Ratio | 0.91 | 0.86 | 1.04 | 1.54 | 1.79 | 1.45 |
| (95% CI) | (0.63, 1.33) | (0.55, 1.34) | (0.56, 1.96) | (0.91, 2.60) | (0.72, 4.46) | (0.79, 2.66) | |
| Obesity (BMI >95th percentile) at 6–10 years old | |||||||
| Maternal first trimester BMI <25 kg/m2 | Odds Ratio | 1.18 | 1.50 | 0.00± | 0.84 | 0.88 | 0.81 |
| (95% CI) | (0.64, 2.17) | (0.80, 2.80) | -- | (0.20, 3.53) | (0.15, 5.26) | (0.09, 7.26) | |
| Maternal first trimester BMI ≥ 25 kg/m2 | Odds Ratio | 0.87 | 0.87 | 0.87 | 1.60 | 1.66 | 1.57 |
| (95% CI) | (0.57, 1.33) | (0.52, 1.45) | (0.43, 1.77) | (0.93, 2.75) | (0.65, 4.27) | (0.84, 2.96) | |
| Overweight/Obesity (BMI >85th percentile) at 11–18 years old | |||||||
| Maternal first trimester BMI <25 kg/m2 | Odds Ratio | 1.06 | 0.96 | 1.47 | 1.49 | 0.77 | 1.86 |
| (95% CI) | (0.65, 1.73) | (0.56, 1.66) | (0.60, 3.63) | (0.50, 4.38) | (0.05, 13.02) | (0.61, 5.68) | |
| Maternal first trimester BMI ≥ 25 kg/m2 | Odds Ratio | 1.12 | 1.13 | 1.11 | 1.05 | 0.74 | 1.23 |
| (95% CI) | (0.68, 1.84) | (0.64, 1.99) | (0.45, 2.73) | (0.40, 2.77) | (0.13, 4.19) | (0.40, 3.79) | |
| Obesity (BMI >95th percentile) at 11–18 years old | |||||||
| Maternal first trimester BMI <25 kg/m2 | Odds Ratio | 1.48 | 0.91 | 4.55 | 1.43 | --† | 2.53 |
| (95% CI) | (0.71, 3.10) | (0.36, 2.31) | (1.61, 12.82) | (0.27, 7.74) | -- | (0.56, 11.38) | |
| Maternal first trimester BMI ≥ 25 kg/m2 | Odds Ratio | 0.84 | 0.74 | 1.11 | 0.52 | 0.27 | 0.72 |
| (95% CI) | (0.46, 1.52) | (0.36, 1.53) | (0.43, 2.92) | (0.15, 1.81) | (0.03, 2.81) | (0.17, 2.94) | |
In the children 6–11 years old, there were no children with obesity born to pregnancies with GGI and 1 abnormal OGTT value
Abbreviations: NGT: normal glucose tolerance, GDM: gestational diabetes.
Maternal first trimester BMI ≥25 kg/m2:
In children 2–5 years old, there were 7047 in the NGT group, 1,247 in the GGI group (851 were in the group with 0 abnormal OGTT values and 396 in the group with 1 abnormal OGTT value), and 553 in the GDM group (225 were CC GDM and 328 were NDDG GDM).
In children 6–11 years old, there were 801 in the NGT group, 125 in the GGI group (87 were in the group with 0 abnormal OGTT values and 38 in the group with 1 abnormal OGTT value), and 70 in the GDM group (20 were CC GDM and 50 were NDDG GDM).
In children and adolescents 12–18 years old, there were 453 in the NGT group, 76 in the GGI group (57 were in the group with 0 abnormal OGTT values group and 19 in the group with 1 abnormal OGTT value), and 24 in the GDM group (7 were CC GDM and 17 were NDDG GDM)
Maternal first trimester BMI <25 kg/m2:
In children 2–5 years old, there were 8,335 in the NGT group, 1,073 in the GGI group (849 were in the group with 0 abnormal OGTT values and 224 in the group with 1 abnormal OGTT value), and 244 in the GDM group (90 were CC GDM and 154 were NDDG GDM).
In children 6–11 years old, there were 1,123 in the NGT group, 143 in the GGI group (111 were in the group with 0 abnormal OGTT values and 32 in the group with 1 abnormal OGTT value), and 26 in the GDM group (13 were CC GDM and 13 were NDDG GDM).
In children and adolescents 12–18 years old, there were 778 in the NGT group, 112 in the GGI group (87 were in the group with 0 abnormal OGTT values and 25 in the group with 1 abnormal OGTT value), and 19 in the GDM group (6 were CC GDM and 13 were NDDG GDM)
Discussion
In this retrospective study of 27,876 children and adolescents exposed to different degrees of glucose intolerance in utero, we found that the proportion and the odds of overweight/obesity and obesity increased with age and worsening glucose exposure categories. The association between GGI and risk of childhood overweight and obesity as compared to the risk in normal glucose tolerance was attenuated by adjustment for maternal BMI which could reflect genetic, familial and environmental factors shared by both the mother and child. However, for older children and adolescents 11–18-year-olds, the risk of long-term obesity was increased in the groups with more severe hyperglycemia even after adjusting for maternal BMI.
Hyperglycemia in pregnancy is associated with perinatal adverse outcomes including large for gestational age and macrosomia, neonatal hypoglycemia, cesarean delivery, shoulder dystocia and clavicular fracture.5,6 The present study sought to evaluate the relationship between different degrees of gestational hyperglycemia and risk of overweight and obesity across childhood and adolescence. In accordance with other studies, the magnitude of this association was greater in older age groups10,23 and may strengthen after puberty.4 Prior evidence has been most consistent in offspring exposed to GDM in pregnancy.7,24,25 In this study, we were able to demonstrate an association between GGI and obesity in older offspring, even after adjusting for maternal BMI. Consistent with observational follow-up of a randomized controlled trial that did not show a difference in BMI at age 4–5 years according to treatment for GDM, we did not find a difference in effect size for untreated versus treated GDM across age categories. It is worth noting that there could be life-style related factors that contributed to the relationship we observed; data on lifestyle factors were not available in our study cohort.
A plausible physiological explanation for our findings is that fetal programming has long range effects on the risk of obesity that contributes to the intergenerational presence of obesity.26 Among Pima Indians, there is an increased risk of obesity in siblings born after a diagnosis of type 2 diabetes in the mother but not the father, with the average BMI of the exposed siblings surpassing the exposed siblings starting at the age of 9 and older.16 Although genetics plays an important role, this study suggests an effect of exposure to hyperglycemia in utero that becomes evident in older children and adolescents. Cho et al. found the association between GDM exposure in utero and BMI z-scores, sum of skinfolds, insulin resistance, leptin z-score and leptin/adiponectin ratio z-scores was stronger in adolescence compared to younger offspring.23 They did not find a similar association among GGI pregnancies, which may have been due to a smaller sample than our own. In our larger cohort, the risk of obesity follows the same pattern of increased prevalence with age for both GGI with 1 abnormal value and GDM exposures. These findings suggest that hyperglycemia in pregnancy may impact the risk of obesity over the life course possibly by triggering disease at the time of a physiologic increase in insulin resistance at puberty.27 Hormones such a leptin and insulin play an important role in the signaling pathway that results in puberty onset.28,29 Epigenetic programming of leptin regulation that occurs in utero can emerge as an obesity phenotype after several years30,31
Our study has both strengths and limitations. This large cohort allows for the long-term analysis of children following an exposure to hyperglycemia in pregnancy across different age groups. Further, our cohort has high generalizability given the representation of different socioeconomic groups and racial and ethnic backgrounds. As with other real-life studies based on electronic medical record data, the information is limited by data collection in clinical encounters. Although we used race/ethnicity, insurance, and marital status to control for social variables and maternal BMI to partially account for genetic impact, we have limited data on environmental factors and other unobserved confounders. Family history of diabetes and history of breastfeeding were not included in multivariate models due to limited or unavailable data. Although individuals with pre-existing diabetes would not have undergone testing in pregnancy and would not have been included in the study, there may be a minority of individuals that had undiagnosed pre-existing diabetes included in our study cohort. There are limitations inherent to observational studies that limit our ability to determine a causal relationship. Finally, our findings are limited by the number of pregnancies that we were unable to link to childhood demographic data if the child data was not available within our health system (excluded participants were more likely to be married/partnered, non-Hispanic White, and have private insurance).
In conclusion, this study shows an association between different degrees of gestational hyperglycemia and risk of childhood and adolescent obesity that is more prominent in older offspring, even after adjusting for maternal BMI. Trajectory studies would be helpful in determining the pattern of weight gain in these children by allowing us to observe weight gain changes throughout childhood using BMI growth curves which are used clinically in all pediatric practices. Further, maternal BMI influences the risk of offspring obesity. In addition to our current glucose centric approach, alternative interventions during pregnancy may be necessary to impact long-term health outcomes in hyperglycemia-exposed offspring at risk for developing obesity.
Financial support:
J.M. was supported by a T32 training grant for Endocrinology (T32DK007028-47S1). The MGH Maternal Health Cohort was supported by the MGH Claflin Distinguished Scholar Award and the MGH Physician-Scientist Development Award. C.E.P. was supported by NIDDK K26DK138346.
Disclosures/Potential Conflicts of Interest:
C.E.P. has received fees and royalties from Mediflix and UpToDate (Wolters Kluwer), respectively, for presentations and articles related to diabetes over which she had full control of content. C.E.P. receives research support from Dexcom through her institution for an unrelated project. The other authors report no conflicts of interest.
Footnotes
Portions of the study findings were presented in abstract/poster form at the American Diabetes Association 2023 and Pediatric Endocrine Society 2024 Scientific Sessions.
Data Availability:
Restrictions apply to the availability of some or all data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.
References
- 1.Stierman B, Afful J, Carroll MD, Chen T, Davy O, Fink S, et al. Prevalence of obesity among youth aged 2–19 years, by sex and age: United States, 2015–2016. National Health and Nutrition Examination Survey 2017–March 2020 Prepandemic Data Files Development of Files and Prevalence Estimates for Selected Health Outcomes. National Center for Health Statistics (US). 2021; [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ward ZJ, Long MW, Resch SC, Giles CM, Cradock AL, Gortmaker SL. Simulation of Growth Trajectories of Childhood Obesity into Adulthood. N Engl J Med. 2017. Nov 30;377(22):2145–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Shah RD, Braffett BH, Tryggestad JB, Hughan KS, Dhaliwal R, Nadeau KJ, et al. Cardiovascular Risk Factor Progression in Adolescents and Young Adults with Youth-Onset Type 2 Diabetes. J Diabetes Complications. 2022. Mar;36(3):108123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Dabelea D, Crume T. Maternal Environment and the Transgenerational Cycle of Obesity and Diabetes. Diabetes. 2011. Jul;60(7):1849–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Roeckner JT, Sanchez-Ramos L, Jijon-Knupp R, Kaunitz AM. Single abnormal value on 3-hour oral glucose tolerance test during pregnancy is associated with adverse maternal and neonatal outcomes: a systematic review and metaanalysis. American Journal of Obstetrics and Gynecology. 2016. Sep;215(3):287–97. [DOI] [PubMed] [Google Scholar]
- 6.Maya J, Selen DJ, Thaweethai T, Hsu S, Godbole D, Schulte CCM, et al. Gestational Glucose Intolerance and Birth Weight–Related Complications. Obstetrics & Gynecology. 2023. Sep;142(3):594–602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 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. Dec 1;40(Supplement_2):121–5. [DOI] [PubMed] [Google Scholar]
- 8.Kawasaki M, Arata N, Miyazaki C, Mori R, Kikuchi T, Ogawa Y, et al. Obesity and abnormal glucose tolerance in offspring of diabetic mothers: A systematic review and meta-analysis. PLoS One. 2018. Jan 12;13(1):e0190676. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Landon MB, Mele L, Varner MW, Casey BM, Reddy UM, Wapner RJ, et al. The relationship of maternal glycemia to childhood obesity and metabolic dysfunction‡. J Matern Fetal Neonatal Med. 2020. Jan;33(1):33–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bendor CD, Bardugo A, Rotem RS, Derazne E, Gerstein HC, Tzur D, Pinhas-Hamiel O, Tsur AM, Cukierman-Yaffe T, Lebenthal Y, Afek A, Chodick G, and Twig G, Afek Arnon, Chodick Gabriel,4,11 and. Glucose Intolerance in Pregnancy and Offspring Obesity in Late Adolescence. Diabetes Care. 2022. Jul;45(7):1540–8. [DOI] [PubMed] [Google Scholar]
- 11.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. May;33(5):964–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hsu S, Selen DJ, James K, Li S, Camargo CA, Kaimal A, et al. Assessment of the Validity of Administrative Data for Gestational Diabetes Ascertainment. American Journal of Obstetrics & Gynecology MFM. 2023. Feb;5(2):100814. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.ACOG Practice Bulletin No. 190: Gestational Diabetes Mellitus. Obstet Gynecol. 2018. Feb;131(2):e49–64. [DOI] [PubMed] [Google Scholar]
- 14.Carpenter MW, Coustan DR. Criteria for screening tests for gestational diabetes. Am J Obstet Gynecol. 1982. Dec 1;144(7):768–73. [DOI] [PubMed] [Google Scholar]
- 15.National Diabetes Data Group. Classification and Diagnosis of Diabetes Mellitus and Other Categories of Glucose Intolerance. Diabetes. 1979. Dec 1;28(12):1039–57. [DOI] [PubMed] [Google Scholar]
- 16.Hampl SE, Hassink SG, Skinner AC, Armstrong SC, Barlow SE, Bolling CF, et al. Clinical Practice Guideline for the Evaluation and Treatment of Children and Adolescents With Obesity. Pediatrics. 2023. Feb 1;151(2):e2022060640. [DOI] [PubMed] [Google Scholar]
- 17.Centers for Disease Control and Prevention, National Center for Health Statistics. Clinical growth charts. Available at: https://www.cdc.gov/growthcharts/clinical_charts.htm. Accessed October 5, 2024.
- 18.Selen DJ, Thaweethai T, Schulte C, Hsu S, He W, James K, et al. Gestational glucose intolerance and risk of future diabetes. Diabetes Care. 2023. Jan;46:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.R Core Team (2023). _R: A Language and Environment for Statistical Computing_. R Foundation for Statistical Computing, Vienna, Austria. <https://www.R-project.org/>. [Google Scholar]
- 20.Højsgaard S, Halekoh U, Yan J. The R Package geepack for Generalized Estimating Equations. Journal of Statistical Software. 2006;15(2):1–11. [Google Scholar]
- 21.Yan J, Fine J. Estimating equations for association structures. Stat Med. 2004. Mar 30;23(6):859–74; discussion 875–877,879–880. [DOI] [PubMed] [Google Scholar]
- 22.Yan J. geepack: Yet Another Package for Generalized Estimating Equations. Vol. 2/3. R-News; 2002. 12–14 p. [Google Scholar]
- 23.Cho S, Rifas-Shiman SL, Harnois-Leblanc S, Aris IM, Oken E, Hivert MF. Associations of Abnormal Maternal Glucose Regulation in Pregnancy with Offspring Adiposity, Insulin Resistance, and Adipokine Markers during Childhood and Adolescence. The Journal of Pediatrics. 2024. May;114100. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Dabelea D, Hanson RL, Lindsay RS, Pettitt DJ, Imperatore G, Gabir MM, et al. Intrauterine exposure to diabetes conveys risks for type 2 diabetes and obesity: a study of discordant sibships. Diabetes. 2000. Dec 1;49(12):2208–11. [DOI] [PubMed] [Google Scholar]
- 25.Pettitt DJ, Knowler WC, Bennett PH, Aleck KA, Baird HR. Obesity in Offspring of Diabetic Pima Indian Women Despite Normal Birth Weight. Diabetes Care. 1987. Jan 1;10(1):76–80. [DOI] [PubMed] [Google Scholar]
- 26.Dabelea D The Predisposition to Obesity and Diabetes in Offspring of Diabetic Mothers. Diabetes Care. 2007. Jul 1;30(Supplement_2):S169–74. [DOI] [PubMed] [Google Scholar]
- 27.Amiel SA, Sherwin RS, Simonson DC, Lauritano AA, Tamborlane WV. Impaired Insulin Action in Puberty. N Engl J Med. 1986. Jul 24;315(4):215–9. [DOI] [PubMed] [Google Scholar]
- 28.Shashaj B, Luciano R, Contoli B, Morino GS, Spreghini MR, Rustico C, et al. Reference ranges of HOMA-IR in normal-weight and obese young Caucasians. Acta Diabetol. 2016. Apr;53(2):251–60. [DOI] [PubMed] [Google Scholar]
- 29.Roddick C, Harris M, Hofman PL. The Metabolic Programming of Pubertal Onset. Clinical Endocrinology. 2025. May;102(5):526–38. [DOI] [PubMed] [Google Scholar]
- 30.Gagné-Ouellet V, Breton E, Thibeault K, Fortin CA, Cardenas A, Guérin R, et al. Mediation Analysis Supports a Causal Relationship between Maternal Hyperglycemia and Placental DNA Methylation Variations at the Leptin Gene Locus and Cord Blood Leptin Levels. Int J Mol Sci. 2020. Jan 3;21(1):329. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Allard C, Desgagné V, Patenaude J, Lacroix M, Guillemette L, Battista M, et al. Mendelian randomization supports causality between maternal hyperglycemia and epigenetic regulation of leptin gene in newborns. Epigenetics. 2015. Mar 24;10(4):342–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Oken E, Kleinman KP, Rich-Edwards J, Gillman MW. A nearly continuous measure of birth weight for gestational age using a United States national reference. BMC Pediatr. 2003. Dec;3(1):6. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Restrictions apply to the availability of some or all data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.