Introduction

Gestational diabetes mellitus (GDM) is defined as glucose intolerance with onset or first recognition during pregnancy. GDM is characterized by pancreatic beta cell function being insufficient to meet the body’s insulin need, usually as a result of insulin resistance (IR) during pregnancy [1]. The established risk factors for GDM are high maternal age and parity, pre-pregnancy overweight/obese, previous delivery of macrocosmic newborns, and family history of diabetes [2]. Research in the past decade has identified a few lifestyle factors that are associated with GDM risk, including a Western dietary pattern, high consumption of red meat and sugary drinks, low consumption of dietary fiber, low glycemic index diet, and physical inactivity before pregnancy [3]. Although GDM has been recognized as one of the most common pregnancy complications, its epidemiology has not been well investigated. One of the reasons is the lack of global consensus on the diagnosis of GDM. Recent data have shown a substantial rise in the prevalence of GDM among women of various ethnic/racial backgrounds [4–8], and would likely further increase as the obesity rate is continually increasing globally. The purpose of the current study is to summarize recent studies about GDM diagnosis, prevalence, and long-term impact on women’s later life and their offspring’s health, and to provide an updated review to guide future research.

Diagnosis of GDM

Diagnosis of GDM is commonly based on the results of an oral glucose tolerance test (OGTT) during pregnancy; however, there is currently no global consensus on the diagnostic criteria. A variety of diagnostic criteria have been used in different studies over time and across regions, including criteria from the National Diabetes Data Group (1997) [9], Carpenter and Coustan (1982, also recommended by the fourth international workshop-conference of GDM, 1998) [10], the World Health Organization (WHO 1985 and 1999) [11, 12], the European Association for the Study of Diabetes (EASD, 1991) [13], American Diabetes Association (ADA 2004, 2011, and 2014) [14, 15, 16•], the International Association of Diabetes and Pregnancy Study Groups (IADPSG, 2010) [17•], and the US National Institutes of Health (NIH, 2013) [18]. There is a general agreement that the OGTT tests should be conducted at gestational weeks 24–28; however, the disagreement is whether to use the 100-g 3-h OGTT or 75-g 2-h OGTT, and different groups propose different cut-offs for the positive results. Table 1 provides descriptions of the diagnostic criteria that have been applied in different time periods and studies. It is important to know these diagnostic criteria because they have significant impact on the prevalence estimates at different times and in different countries. In a 2005 study of the agreement between several international diagnostic criteria, the prevalence of GDM in the same population ranged from 8 to 25 % depending on the criteria used (Fig. 1) [19], and the ADA and WHO criteria showed only a moderate correlation (kappa statistic = 0.51). In the current review, these inconsistencies are important to consider when comparing prevalence data across studies that used different diagnostic criteria.

Table 1 Gestational diabetes mellitus diagnostic criteria (24–28 gestational weeks)
Full size table
Fig. 1
figure 1

The Different Diagnostic Criteria on GDM Prevalence in One Population of Pregnant Women. CDA Canadian Diabetes Association, EASD European Association for the Study of Diabetes, ADA American Diabetes Association, NZSSD New Zealand Society for the Study of Diabetes, WHO World Health Organization, ADIPS Australian Diabetes Pregnancy Society. Reference [19]

Full size image

In 2004, the ADA proposed that GDM can be detected with either of two strategies: (1) “one-step” approach of 75-g OGTT or (2) “two-step” approach with a 50-g screen followed by one 100-g OGTT for those who screened positive. The same cut-off points are used in both tests (fasting ≥ 5.3 mmol/l; 1-h ≥ 10 mmol/l; and 2-h ≥ 7.8 mmol/l). However, it is required that two of three abnormal values are needed for diagnosis [14]. After deliberations in 2008 to2009, the IADPSG recommended testing women with risk factors for undiagnosed type II diabetes (T2D) at their initial prenatal visit. Women who are diagnosed with diabetes in the first trimester would be classified as T2D. GDM is diabetes diagnosed based on 75-g 2-h OGTT in the second or third trimester of pregnancy. This recommendation has been made based on findings from the Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study and other works that examined associations of maternal glycemia with perinatal and long-term outcomes in offspring, and has been endorsed by multiple obstetrical and diabetes organizations [17•]. In 2011 Standard of Care, the ADA for the for the first time recommended that all pregnant women not known have prior diabetes undergo a 75-g OGTT at 24–28 weeks of gestation based on the recommendation of the IDAPSG [15]. In 2013, the NIH called a panel with 15 representatives from multiple fields (i.e., obstetrics/gynecology, maternal-fetal medicine, pediatrics, diabetes research, and biostatistics) to discuss the diagnostic criteria of GDM [18]. The panel recommended the two-step approach. These approaches differed on whether a 1-h sample should be included, whether two abnormal values are required, and the diagnostic cut-off values that were used. Given the fact that there are data to support both approaches, ADA recommended both two options for the diagnosis of GDM in 2014 [16•].

Prevalence of GDM

Since there is no global consensus on the diagnostic criteria for GDM, the prevalence of GDM ranged from 1.0 to 14.2 % of all pregnancies depending on diagnostic criteria and the study population (based on our literature search of Medline publications with either population-based studies with sample size ≥500 or hospital-based studies with sample size ≥1000 and at least 70 % of population being screened for GDM) As shown in Table 2, the Southeast Asia region consistently reported the highest GDM prevalence with a median of 5.4 % (range of 3–14.2 %), followed by Eastern Mediterranean countries with a median of 4.75 % (range of 1.9–13.7 %). America, Africa, and Western Pacific appeared to have similar GDM prevalence with a median of 3.7 % across the three regions. Europe had the lowest GDM prevalence among all the WHO regions. The low prevalence was evident across all European countries (range 1.2–3.1 %) except for Italy. A population-based study in Sardinia, Italy, provided the largest GDM prevalence (22.3 %) of all the included studies [20]. A second large population-based study from a mainland Italian city, Pisa, also reported a GDM prevalence of 8.7 %.

Table 2 Worldwide prevalence of gestational diabetes mellitus
Full size table

The prevalence data in the USA over time are presented in Fig. 2. The study that had the largest time span, 15 years, showed the largest difference in prevalence, 2.6 % in 1990 and 7.5 % in 2005 [21]. A recent study using data from the Pregnancy Risk Assessment Monitoring System (PRAMS) reported that the GDM prevalence (included 21 states that participated in PRAMS from 2007–2010) was 8.1 % in 2007–2008 and 8.5 % in 2009–2010 [22•]. Another study using the Agency for Healthcare Research and Quality’s (AHRQ) National and State Inpatient Database (included 12 states) found that the GDM prevalence increased from 3.71 % in 2000 to 5.77 % in 2010 among all hospital deliveries [23]. Although other factors such as the diagnostic criteria, the geographic location, and the changing demographics of the study population must be acknowledged when comparing these studies and the actual numeric prevalence estimates, there was a clear trend on the increase in GDM prevalence in the USA from the 1990s to 2010s.

Fig. 2
figure 2

Trends of GDM in the USA from the 1990s to the 2010s. References [4, 5, 21, 91, 108-110]

Full size image

US prevalence data showed a clear difference across the races/ethnicities that was consistent throughout all the studies (Table 3). Whites had the lowest prevalence of GDM followed by Blacks, then Hispanics, and Asian Americans had the highest prevalence of GDM. This disparity was consistent across the almost 20-year time span of the studies included. As the data in Table 3 are displayed in ascending order by year of the study, we can see an increase in GDM prevalence for each race/ethnicity group over time and the highest increase was observed in Hispanics [23].

Table 3 Prevalence (%) of GDM in the USA by race/ethnicity
Full size table

The geographic distribution of gestational diabetes in the USA is presented in a map as Fig. 3. Prevalence data was available for 21 states (with either population-based studies with sample size ≥500 or hospital-based studies with sample size ≥1000 and at least 70 % of population being screened for GDM). The states with the highest GDM prevalence, between 5.6 and 8.0 % of pregnancies, are Texas, California, Hawaii, and Alaska. This geographic difference of GDM may be largely explained by the racial/ethnic composition in each state. According to the 2000 census data, half of all Hispanics in the USA lived in just two states, Texas and California. California also had a substantial Asian population. Since Asian and Hispanic populations had a higher prevalence of GDM than other ethnicities, this may explain the high prevalence found in Texas and California. Hawaiian and Alaskan natives have also been implicated as high-risk ethnicities in several studies [24, 25]. Despite the agreement between the geographic and the ethnic disparity, more data should be collected for all of the states in order to further study regional trends and differences.

Fig. 3
figure 3

State-level geographic disparities of GDM in the USA. References for GDM prevalence by state: AK [25]; CA [106]; CO [107]; CN [111]; GA [112]; HI [113]; IL [114]; IN [115]; MT [116]; NE [117]; NY [105]; OK [118]; OR [119]; PA [120]; TX [121]; UT [122]; WV [123]

Full size image

Long-Term Maternal Outcomes of GDM

Women with GDM are at greater risk for other pregnancy complications such as maternal hypertensive disorders and cesarean delivery [26]. Although most women with GDM regain normal glucose tolerance after delivery, many of them are featured with metabolic disorders in postpartum and later life. Women with GDM are at increased risk of developing impaired glucose tolerance (IGT) and T2D in their later life. On average, the risk of development of T2D is 7.4 times greater in GDM women than non-GDM women [27••]. Such increased risk of T2D among women with GDM has been documented in different populations and countries [28]. In a large Canadian population-based study, the probability of developing T2D among women with GDM was 3.7 % at 9 months and 18.9 % at 9 years after delivery. In contrast, the rate was only 2.0 % at 9 years for women who did not have GDM [29]. Another study in Australia also found risk of development of diabetes in GDM women increased with time from the index pregnancy: the cumulative incidence of diabetes is 2.6 % in 2 years, 8.1 % in 5 years, 17.3 % in 10 years, and 25.8 % in 15 years [30]. Moreover, recent data have indicated that women with a history of GDM might be at an increased risk of cardiovascular diseases (CVD) independent of T2D or obesity. Several studies have reported an association between GDM and CVD risk factors, including obesity, high blood pressure, abnormal plasma lipid levels, metabolic disorder, inflammation, and endothelial dysfunction [31–39]. In addition, other nontraditional CVD risk biomarkers, such as higher level of C-reactive protein (CRP), E-selectin, fibrinogen, plasminogen activator inhibitor (PAI)-1, and lower circulation level of adiponectin, have been fund in women with a history of GDM [37, 38, 40–43]. Moreover, recent studies have directly linked a history of GDM with a CVD clinical event. A US study reported that among women with a family history of T2D, those with GDM were more likely to have a CVD event (odds ratio (OR) = 1.85; P = 0.005), independent of T2D [44]. The higher risk of development of CVD among women with a history of GDM was also reported by a larger study in Canada with a median follow-up of 11.5 years [45]. A recent publication from CARDIA (The Coronary Artery Risk Development in Young Adults) study found that the history of GDM was associated with atherosclerosis (measured by common carotid intima-media thickness), independent of pre-pregnancy obesity [46]. Taken together, emerging evidence tends to suggest that women with GDM represent a population who are more likely to develop CVD at a relative younger age.

Long-Term Outcomes in the Offspring of Women with GDM

Theoretically, exposure to a diabetic intrauterine environment results in excess fetal growth. Biologically, maternal glucose can cross the placenta but maternal insulin does not. To respond to this increase in glucose load, the fetal pancreas increases the insulin production, which in turn promotes fetal growth and adiposity development [47]. There is a growing in the literature that reports a positive association between maternal hyperglycemia and higher risk of development of obesity and diabetes in offspring. A few studies also found a direct link between maternal diabetes and offspring CVD risk factors, including blood pressure [48–51] and plasma lipids [49, 50]. One recent study by Kaiser Permanente South California found that exposure to GDM by 26 weeks’ gestation was associated with offspring autism [52]. Among those health outcomes, the strong evidence was found for offspring obesity and T2D.

Early evidence from Pima Indian data indicated that offspring of women with preexisting diabetes or GDM were heavier at birth and had an increased risk of obesity during childhood and young adulthood (5–19 years) [53–55]. These findings were confirmed by the Diabetes in Pregnancy Study at Chicago in which children born to women with diabetes during pregnancy (both pregestational diabetes and GDM) were on average 30 % heavier than expected for their height at age at 8 years [48]. While some studies did not find such a positive association between maternal diabetes and offspring obesity, a recent systematic review and meta-analysis of nine studies indicated that maternal diabetes was associated with increased offspring BMI [56]. Although most of the studies used the BMI to define overweight or obesity in the offspring, Catalano et al. showed that newborns of women with GDM had 20 % greater fat mass (measured by body electrical conductivity) compared with infants of women with normal glucose tolerance during pregnancy. Maternal fasting glucose was the strongest predictor of fat mass in newborns of women of GDM [57].

Several studies also found a positive association between maternal diabetes with insulin resistance, impaired glucose tolerance (IGT) and elevated risk of T2D in offspring. In the Diabetes in Pregnancy Study, the offspring of diabetic mothers had higher 2-h glucose and insulin concentrations when compared with the controls of nondiabetic mothers at age 10–16 years [58]. In the Pima Indian Study, there was a significantly higher prevalence of T2D in the offspring of diabetic mothers at all ages, ranging from 5 to 34 years, than in the offspring of prediabetic and nondiabetic mothers [59]. In addition, the insulin secretion was impaired in Pima adults whose mothers were diabetic during pregnancy than in those whose mothers developed diabetes after their births [60].

Although studies conducted from different populations consistently showed that maternal diabetes is associated with an increased risk of having obesity and T2D in offspring in both childhood and young adulthood, there are still questions and debates as to (1) whether such association is independent of the type of diabetes during pregnancy (i.e., pregestational type I or II diabetes or GDM), (2) whether the association of maternal diabetes and offspring diabetes is because of the exposure to a diabetic intrauterine environment during pregnancy in addition to genetic transmission of diabetes susceptibility between mother and their offspring, and (3) whether such associations are independent of maternal obesity (pre-pregnancy BMI or gestational weight gain). In the following sections, we review recent studies that provide data to answer these questions.

While most earlier studies did not differentiate the type of maternal diabetes in their analyses, recent evidence tends to suggest that the long-term consequence of exposure to maternal diabetes is not related to the type of diabetes during pregnancy. In the Diabetes in Pregnancy Study, the occurrence of IGT in offspring was associated with maternal hyperglycemic but not the type of diabetes in the mothers [58]. In German studies, the occurrences of IGT and obesity at 1–4 years and 5–9 years were similar in offspring of mothers with preexisting type I diabetes and in those of mothers with GDM [61, 62].

Studies that investigate the role of diabetic intrauterine environment on offspring obesity or diabetes are limited, but results are consistent from two studies in different populations. In the Pima Indians, the prevalence of T2D was compared between siblings born before and after their mother developed diabetes [59]. The risk of diabetes was significantly higher in siblings born after the mother developed diabetes as compared with those born before the mother’s diagnosis of diabetes (OR = 3.7, P = 0.02). Additionally, among 183 siblings without diabetes, the mean BMI was 2.6 kg/m [2] higher (P = 0.003) in the offspring of diabetic than in offspring of nondiabetic pregnancies. Since siblings are considered to have a similar risk of carrying genetic family environmental factors for diabetes, the different risks between siblings likely reflects the influence of intrauterine diabetes environments. A recent sibling study in Swedish men also found that maternal diabetes was associated with greater mean BMI of their son at age of 18 years [63••]. The BMI of men whose mother had GDM was 0.94 kg/m [2] greater than in their brothers born before their mothers were diagnosed with diabetes, suggesting this association is importantly driven by an intrauterine mechanism.

Recently, there is a debate as to whether the association between maternal diabetes and offspring obesity or glucose intolerance is independent of maternal obesity during pregnancy. Results from a birth cohort in Finland indicated that maternal pre-pregnancy overweight is associated with a higher risk of offspring obesity (OR = 2.56; P < 0.001) at age of 16 years, even among those mothers with a normal glucose during the pregnancy. However, in offspring with pre-pregnancy normal weight, the risk of obesity was not increased by prenatal exposure to GDM (OR = 0.73, P > 0.05) [64]. Data from the Avon Longitudinal Study of Parents and Children (ALSPAC) in the UK suggested GDM was associated with greater BMI and fat mass of offspring at 9–11 years of age. Although the association attenuated with adjustment of maternal early pregnancy BMI, the positive association remained [65]. This finding is also supported by two meta-analyses [56, 66]. A large population-based study in the USA using Kaiser Permanente data reported that the risk of offspring obesity was positively associated with hyperglycemia in pregnancy, independent of gestational weight gain [67]. A recent large prospective study in Swedish men also showed that maternal diabetes (both pregestational diabetes and GDM) was associated with greater offspring BMI in early adulthood, independent of maternal early pregnancy BMI [63]. Results from the SEARCH Case-Control study indicated that exposure to maternal diabetes (both pregestational diabetes and GDM: OR = 5.7) and exposure to maternal overweight/obesity (measured by pre-pregnancy BMI ≥25 kg/m [2]: OR = 2.8) were independently associated with T2D in youth (10–22 years) across different race/ethnicity groups (i.e., White, African-Americans, and Hispanics). For those youth who had intrauterine exposure to both maternal diabetes and obesity with the OR of T2D as 19.2 suggests an interactive effect [68]. Therefore, available evidence on the interrelationship of maternal diabetes, maternal obesity, and offspring risk of obesity/diabetes is inconsistent. One possible explanation for those studies failing to show the association is a possible attenuation of the risk due to the treatment of maternal diabetes during pregnancy. Another explanation could be that studies used different methods and criteria for the diagnosis of diabetes during pregnancy. Moreover, different studies have examined offspring obesity at various ages. Although most studies found a positive association between maternal diabetes and offspring obesity, it is still unclear at what age this association becomes apparent. Findings from Project Viva indicated that GDM predicted a slower weight gain in the first 6 months of life [69]. This finding was confirmed by longitudinal results from the Exploring Perinatal Outcomes among Children (EPOCH) Study in which the growth trajectory from birth through 13 years of age was compared between children with and without exposure to maternal diabetes. The overall BMI growth curve (adjusted for sex and race) was not significantly different for exposed and unexposed children from birth through 26 months of age (P = 0.48). However, the overall BMI curve from 27 months of age through 13 years differed significantly by exposure status (P = 0.008), indicating that exposure to maternal diabetes accelerates offspring BMI growth in late childhood thus increasing the long-term obesity risk [70].

In summary, current evidence indicates that maternal diabetes has long-term effects on the risk of diabetes and obesity in the offspring. The effects of exposure to maternal diabetes during pregnancy are similar regardless of the type of diabetes (i.e., preexisting type I or II diabetes or GDM). The higher risk of obesity and T2D in the offspring of diabetic mothers is likely due to the intrauterine environment which is independent of shared socioeconomic, lifestyle, and genetic factors within families. It is unclear how pre-pregnancy obesity and gestational weight gain interplay with maternal diabetes to influence the long-term health outcomes in offspring.

Conclusions

GDM presents a significant risk factor for development of T2D and CVD in women. Therefore, identifying women at high risk of GDM has important public health and clinical significances. The initial risk prevention approach for this population should be targeting a reduction in the risk of T2D. Additional efforts should include managing optimal body weight, blood pressure, blood lipids, and other CVD risk factors in this population. Ideally, such efforts will be considered and integrated into current clinical care in both prenatal and postpartum care. Future investigations will be needed to address how to do risk classification among women with GDM. Children whose mothers had diabetes during pregnancy are at increased risk of having obesity and T2D at a young age. Studies are needed to understand the relative contribution of maternal diabetes and obesity to the long-term effects on offspring risk for obesity and T2D. Future research is also needed to determine whether better glucose control during pregnancy would prevent the long-term consequences for both women and their children.