Approximately 4–6% of pregnancies in the United States are complicated by diabetes mellitus (DM), making it one of the most common serious medical complications of pregnancy [1–5]. The majority of cases (>80%) are diagnosed for the first time during pregnancy (gestational DM); the remaining cases are pregestational DM (type 1 or type 2) [2]. Furthermore, the prevalence of diabetes in pregnancy is increasing in the United States, concurrent with the rising prevalence of obesity and type 2 diabetes in the general population; this increase is not explained by changes in the prevalence of other known maternal risk factors, such as advanced maternal age [6–8].

It is commonly stated that intrauterine exposure to maternal diabetes places offspring at increased risk for long-term adverse outcomes including overweight and obesity. In addition, it has been suggested that infants of women with diabetes should be specifically targeted for obesity prevention interventions. This suggestion raises several issues. Maternal obesity is also a risk factor for offspring overweight and obesity, is associated with maternal diabetes, and is a more prevalent condition than either gestational or pregestational DM [9–12]. It is unclear, therefore, whether it would be most effective to target infants of women with diabetes specifically, or all infants of women with prepregnancy obesity. Second, studies commonly cited as providing support for a causal association between maternal gestational DM and offspring overweight and obesity, including those conducted among Pima Indian women, have design limitations including combining pregestational and gestational DM into one exposure group, examination of exposed individuals only without a nondiabetic control group, and failure to control for important potential confounders [13–15].

In this study, we systematically reviewed studies examining the association between gestational DM and childhood overweight and obesity. We focused on gestational DM because it is the most prevalent form of diabetes in pregnancy. We summarized findings and addressed methodological limitations of previously published studies. We did not calculate a combined single estimate of the findings due to the heterogeneity of methodology and rigor among the studies. Finally, we provided recommendations regarding approaches for future studies.

We identified 1362 potentially relevant studies by searching PubMed; of these, 144 abstracts reported a finding on the relationship between maternal gestational DM and childhood obesity and the full texts of these articles were retrieved for detailed examination (Figure 1). We also reviewed the reference lists of all 144 studies, and identified an additional 48 studies for possible inclusion. After review of these 192 articles, 140 were excluded, either because they did not report the results of a cohort or case-control study or because they clearly did not address an association between maternal gestational DM and childhood obesity; thus, 52 studies were considered further for inclusion. Of these, 37 studies were excluded because gestational and pregestational DM were combined, there was no nondiabetic control group, or BMI was not presented as 85th or 95th percentile or it was presented only as a continuous variable. An additional 3 studies were excluded because later papers reported results on the same study population. A total of 12 studies met the inclusion criteria [17–28].

Six of the studies were conducted in the United States, and one each was conducted in Germany, Brazil, the United Kingdom, Hong Kong, Finland, or Poland (Table 1). We translated one study and contacted three authors. No studies were excluded because of lack of language translation resources. Sources of information on maternal gestational DM diagnosis included maternal self-report, clinical records, or blood glucose measurements. Information used to determine childhood obesity was obtained from parental report, anthropometric measurements obtained as part of a study protocol, hospital records, or clinical databases.

The crude odds ratios for the relationship between gestational DM and childhood overweight or obesity ranged from 0.7 to 6.3 (Figure 2(a)). When the studies were ordered from offspring overweight to obesity, the magnitude of the association did not increase with increasing levels of offspring BMI. If we exclude the lower and upper values, the OR ranges from 1.0 to 2.5. Eight of the twelve studies had significant findings in the crude analysis. In the three of the twelve studies (two of which were among the eight with a significant finding), the authors adjusted for potential confounders when examining the association between gestational DM and childhood obesity compared to a nondiabetic control group (Figure 2(b)). In the first of the three studies, Gillman and colleagues reported a crude odds ratio (OR) of 1.2 (0.9–1.5) and 1.4 (1.1–2.0) for adolescent weight at 85th–95th percentile and >95th percentile, respectively at age 9–14 years old among the offspring of mothers with gestational DM versus no diabetes [19]. These estimates did not change after adjustment for level of child's physical activity, hours per week watching television, and energy intake. However, adjustment for maternal prepregnancy BMI attenuated the associations for both weight at 85th–95th percentile (AOR = 1.0, 95% CI 0.7–1.3) and at >95th percentile (AOR = 1.2, 95% CI 0.8–1.7), and results were no longer significant. In the second study, Lawlor and colleagues reported a crude OR of 1.5 (0.8, 2.9) for weight >85th percentile at age 9–11 years among offspring of mothers with gestational DM, but this positive finding was reversed after adjustment for prepregnancy BMI (AOR = 0.62, 95% CI 0.3, 1.23) [26]. However, the finding was not significant either before or after the adjustment. In the third study, by Hillier and colleagues, the authors compared treated and nontreated women with gestational DM each to women with no diabetes [20]. They did not adjust for prepregnancy BMI but did adjust for maternal age, parity, gestational weight gain, ethnicity, macrosomia, and sex of child. They reported an adjusted OR of 1.9 (1.3, 2.8) for child's weight >85th percentile and 1.8 (1.1, 2.9) for child's weight >95th percentile at age 5–7 years among offspring of mothers likely not treated for gestational DM (those who met criteria for gestational DM using Carpenter and Coustan criteria only and did not meet the National Diabetes Data Group (NDDG) criteria) compared to offspring of women with no diabetes. They found no significant associations between gestational diabetes and childhood weight >85th percentile (adjusted OR = 1.3 (0.8, 1.9) or for child's weight >95th percentile (adjusted OR = 1.38 (0.8, 2.3)) in women likely treated for gestational DM (those who met criteria for gestational DM using the NDDG criteria). In addition, the authors observed a dose-response relationship between maternal glucose concentration in quartiles among women with no diabetes and risk of childhood overweight and obesity, suggesting that elevated maternal blood glucose does affect risk of overweight and obesity in offspring, and that this risk may be reduced with treatment of gestational DM during pregnancy.

In one of the 12 studies, Boerschmann and colleagues did not adjust for any confounders; however, the authors did present results stratified by maternal BMI among offspring of women with gestational DM. They found that among offspring of women with gestational DM, there was an increase in offspring BMI percentile ≥90 as maternal BMI increased. This increase further supports the strength of maternal BMI [27]. None of the 12 studies included information on genetic factors.

To our knowledge, ours is the first study to systematically review all published case-control and cohort studies specifically examining the association between maternal gestational DM and the prevalence of offspring overweight and obesity. We excluded many studies from our review because of methodological limitations, including the use of an exposure group combining pregestational and gestational DM, and the lack of a nondiabetic control group. We found that studies of the effects of maternal gestational DM on offspring overweight and obesity have yielded inconclusive results, which is consistent with a recent review examining maternal diabetes and offspring BMI z-scores, where they found the association between maternal diabetes and offspring BMI to be no longer significant after adjustment for prepregnancy BMI [29].

Women with preexisting diabetes have a substantially increased risk of pregnancy complications including fetal loss, perinatal mortality, and of delivering an infant with congenital anomalies [30]. Although the mechanisms underlying these associations are not completely understood, it has been shown that tight glycemic control in early pregnancy reduces the prevalence of such pregnancy complications [31]. In contrast to women with pregestational DM who are hyperglycemic throughout pregnancy, women with gestational DM typically develop hyperglycemia in the second or third trimester of pregnancy, and pregnancy complications associated with gestational DM are not as severe [32, 33]. Therefore, we cannot assume that pregestational DM and gestational DM have the same effects on fetal development and long-term offspring outcome, and it is important to study associations between these two conditions and offspring outcomes separately.

In addition to timing of exposure, it is important to consider differences in diagnostic cutpoints for gestational DM. Depending on the cutpoints, it may result in the inclusion of women with milder disease in the gestational DM exposed group, which could attenuate the association between gestational DM and offspring overweight and obesity. Therefore, the severity of disease would be important to consider when examining offspring overweight and obesity.

If treatment of gestational DM reduces risk of overweight or obesity in offspring, it may be difficult to consistently detect an association in observational studies. Women with gestational DM are typically instructed to monitor blood sugar and adjust their dietary habits, which may improve long-term offspring outcomes. In only one study did the authors examine treated and nontreated women with gestational DM [20]. They reported significant associations with childhood overweight and with obesity in the nontreated group but not in the treated group. Further evidence that treatment of gestational DM may improve offspring outcome is found in studies of women with high blood glucose concentration, but not sufficiently high for a diagnosis of gestational DM thus not treated. For example, Deierlein et al. found that fetal exposure to maternal glucose concentration in the high-normal range was associated with the development of overweight and obesity in the offspring at age three years, independent of maternal prepregnancy BMI [37].

As previously discussed, most existing observational studies cannot be used to definitively quantify the independent contribution of gestational diabetes to offspring overweight and obesity risk because of unmeasured confounding and other methodological limitations. We recommend the following for future studies.

This study has some limitations. First, we did not include unpublished studies, therefore our results may be affected by publication bias. In addition, we may have missed studies that were not listed in PubMed or referenced in other published studies or reviews. Finally, these studies are not all directly comparable due to discrepancies in the study population, methodology, gestational DM diagnostic criteria, BMI reference population, and ages. However, the present study is the first to systematically review all published case-control and cohort studies examining the association between maternal gestational DM and the prevalence of offspring overweight and obesity.

In conclusion, this review demonstrates that studies of associations between maternal gestational DM on offspring overweight and obesity have yielded inconclusive results. Because maternal obesity is a more prevalent condition than gestational DM and is strongly associated with offspring obesity, we need a better understanding of the relative contributions of maternal obesity and gestational DM to risk before designating infants of women with GDM specifically as targets for obesity prevention efforts. Interventions addressing prepregnancy obesity may have a greater public health impact on childhood overweight and obesity than those targeting offspring of women with gestational DM. A stronger body of evidence is needed to better understand the potential associations between maternal gestational DM and offspring outcomes.