A Clinical Update on Gestational Diabetes Mellitus

Abstract

Graphical Abstract

Open in a separate window

Graphical Abstract

Diabetes in pregnancy was first described in 1824 by Bennewitz in Germany (1), with subsequent case series in the United Kingdom and United States reporting high perinatal mortality rates in women with diabetes in pregnancy (2-4). In 1909, Williams reported arguably the first diagnostic criteria for diabetes in pregnancy in the United States, proposing physiological and pathophysiological thresholds for “transient glycosuria in pregnancy” (5).

In 1964, O’Sullivan and Mahan defined specific diagnostic criteria for gestational diabetes mellitus (GDM) in the United States derived from the 100-g 3-hour oral glucose tolerance test (OGTT) undertaken in the second and third trimester of pregnancy in 752 women (6). GDM was defined as ≥2 venous whole blood glucose values greater than 2 SD above the mean glucose values for pregnancy in their initial cohort. These glucose thresholds were primarily chosen because the resulting GDM prevalence of 2% corresponded to the background population prevalence of diabetes, while the requirement of ≥2 elevated glucose values sought to minimize the risk of preanalytical error (7). These thresholds were validated by their identification of subsequent diabetes up to 8 years postpartum in an additional cohort of 1013 women. Increased perinatal mortality was also observed in women with ≥2 glucose values exceeding the proposed diagnostic criteria (6). In 1965, the World Health Organization (WHO) concurrently recommended that GDM be diagnosed by either a 50- or 100-g OGTT using the 2-hour postload glucose value, but the threshold used was the same as for diagnosing diabetes in the nonpregnant population (8). The WHO continued to diagnose GDM based on glucose thresholds for diabetes in the nonpregnant population (9,10) until its endorsement of the International Association of the Diabetes and Pregnancy Study Groups (IADPSG) diagnostic criteria in 2013 (11).

Since 1973, the screening approach to GDM frequently adopted a 2-step procedure with the 50-g 1-hour glucose challenge test (GCT) followed by the 100-g 3-hour OGTT if the GCT was positive. This was based on data from O’Sullivan et al, which showed that a 2-step diagnostic approach to GDM using the GCT as the initial screening test and a glucose threshold of 7.9 mmol/L (143 mg/dL) was 79% sensitive and 87% specific for diagnosing GDM on the 100-g 3-h OGTT in a cohort of 752 women (12). The rationale for this approach was the efficient identification of women most at risk of GDM.

In 1979, the US National Diabetes Data Group (NDDG) published conversions of the original O’Sullivan and Mahan 100-g 3-hour OGTT diagnostic criteria for GDM, reflecting the transition from venous whole blood glucose to plasma blood glucose analysis (13). These revised criteria were subsequently adopted by the American Diabetes Association (ADA) and internationally (9,14,15). In 1982, Carpenter and Coustan recommended lowering of the NDDG diagnostic criteria, reflecting newer preanalytical enzymatic methods that were more specific for plasma glucose (7,16). They also advised lowering the GCT glucose threshold to 7.5 mmol/L (135 mg/dL) based on their study of 381 women who underwent the 100-g 3-h OGTT after screening positive on the GCT, whereby a GCT glucose threshold ≤ 7.5 mmol/L (135 mg/dL) strongly correlated with a normal OGTT (17). However, in the absence of clear evidence supporting a specific glucose threshold for the GCT, the ADA and the American College of Obstetricians and Gynecologists (ACOG) continued to recommend a screen positive GCT glucose threshold from 7.2 to 7.8 mmol/L (130-140 mg/dL) for GDM (18,19).

The ADA did however recommend the modified Carpenter and Coustan diagnostic glucose thresholds for GDM from 2000 (20), supported by the findings of the Toronto Tri-Hospital Gestational Diabetes Project (21,22). These data demonstrated a positive correlation between increasing maternal hyperglycemia even below the NDDG diagnostic criteria for GDM and risk of obstetric and neonatal complications including preeclampsia, cesarean section, and macrosomia (neonatal birthweight > 4000 g) (21,22). In addition, several large cohort studies showed that women diagnosed (but not treated) with GDM based on the Carpenter and Coustan criteria were at increased risk of perinatal complications including hypertensive disorders of pregnancy, increased birthweight, macrosomia, neonatal hypoglycemia, hyperbilirubinemia, and shoulder dystocia, compared to women diagnosed and treated as GDM by NDDG diagnostic criteria (16,23-25). From 2003 the ADA additionally endorsed the 1-step 75-g 2-hour OGTT for the diagnosis of GDM derived from the modified Carpenter and Coustan fasting, 1- and 2-hour glucose thresholds for the 100-g 3-hour OGTT, particularly for women at high-risk (26). This approach was deemed more cost-effective, albeit less validated, than the 100-g 3-hour OGTT. The use of the modified Carpenter and Coustan thresholds was associated with an almost 50% increase in prevalence of GDM (16,23).

The evolution of diagnostic criteria for GDM illustrates the historic lack of consensus for the diagnosis of GDM, with the presence or absence of disease varying dependent on expert consensus. The underlying rationale for the diagnosis of GDM also shifted over time toward identifying perinatal risk rather than future maternal diabetes risk.

Current GDM Diagnostic Criteria

The seminal Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) study sought to provide an evidence base to guide risk in GDM, and its results were published in 2008 (27). This large, international, prospective, observational study evaluated the relationship between glucose levels on the 75-g 2-hour OGTT performed at 24 to 32 weeks’ gestation (mean 27.8 weeks’ gestation) in over 25 000 pregnant women with the following primary perinatal outcomes: birthweight > 90th percentile for gestational age, primary cesarean section delivery, neonatal hypoglycemia, and cord blood serum C-peptide > 90th centile. Secondary outcomes were preeclampsia, preterm delivery (defined as delivery before 37 weeks’ gestation), shoulder dystocia or birth injury, hyperbilirubinemia, and neonatal intensive care admission. The results showed a continuous positive linear relationship between maternal fasting; 1- and 2-hour plasma glucose levels obtained on the OGTT, below those that were diagnostic of diabetes outside pregnancy; and risk of primary outcomes (27). Notably, there were no specific glucose thresholds at which obstetric and neonatal complications significantly increased.

Based on these findings and supported by trials [the Australian Carbohydrate Intolerance Study in Pregnant Women (ACHOIS) and the Maternal-Fetal Medicine Units Network (MFMU) trial] showing benefit of treatment of more severe and “mild” degrees of maternal hyperglycemia, respectively (28,29), the IADPSG revised its diagnostic criteria for GDM. Despite the lack of a clear diagnostic glucose threshold in HAPO, the consensus of the IADPSG was to define diagnostic thresholds for the fasting, 1- and 2-hour glucose values for the 75-g 2-hour OGTT based on the average glucose values at which the odds of the primary outcomes were 1.75 times the odds of these outcomes occurring at the mean glucose levels for the HAPO cohort (30). The IADPSG consensus was also that only 1 elevated glucose level for the OGTT was required for GDM diagnosis, as each glucose threshold represented broadly comparable level of risk. Thus, the main purpose of the diagnostic criteria for GDM post-HAPO was to define the level of risk associated with increased perinatal complications.

Post-HAPO, there exist several different screening and testing approaches for the diagnosis of GDM internationally. The IADPSG and WHO recommend universal testing of all pregnant women between 24 to 28 weeks’ gestation with the 75-g 2-hour OGTT (11,30). These revised recommendations were largely endorsed by several organizations including the ADA (18), Endocrine Society (31), International Federation of Gynecology and Obstetrics (32), Australasian Diabetes in Pregnancy Association (33), Japan Diabetes Society (34), Ministry of Health of China (35), and the European Board of Gynecology and Obstetrics (36).

The National Institutes of Health did not endorse the IADPSG recommendations, citing the expected increase in prevalence of GDM, cost, and intervention in the context of a lack of evidence for an associated improvement in perinatal outcomes (37). The National Institutes of Health and ACOG continue to recommend a 2-step testing approach, with the initial screening GCT for all women and those who screen positive proceeding to the diagnostic 100-g 3-hour OGTT (19,37). This approach is also endorsed by ADA (18). However, the ACOG’s 2018 guidelines now acknowledge that individual practices and institutions may instead choose to use the IADPSG’s 1-step testing approach and diagnostic criteria if appropriate for their population (19). The UK National Institute for Health and Care Excellence (NICE) guidelines advise a selective screening approach, whereby women with risk factors for GDM are recommended to undergo a diagnostic 75-g 2-hour OGTT at 26 to 28 weeks’ gestation, with diagnostic (fasting or 2-hour) glucose thresholds higher than the IADPSG diagnostic criteria for GDM (38). Several other European bodies also currently recommend selective risk factor-based screening, with only women fulfilling specific high-risk criteria proceeding to a diagnostic OGTT, even if the IADPSG diagnostic criteria for GDM are applied (39,40). The revised IADPSG diagnostic criteria and testing approach to GDM in comparison to other international organizations are summarized in Table 1.

It is important to consider the increase in GDM prevalence associated with the IADPSG diagnostic criteria in the context of the rising background rates of impaired glucose tolerance, type 2 diabetes, and obesity among young adults and women of reproductive age (46,47). For example, almost 18% of HAPO study participants would have met the IADPSG diagnostic thresholds for GDM. By comparison, the rate of prediabetes in US adults aged between 20 and 44 years is >29% (48,49).

Studies in Indian, Israeli, and US cohorts have suggested that the IADPSG testing approach and intervention for GDM is cost-effective based on a combination of delaying future type 2 diabetes and preventing perinatal complications (50-53). For example, a US study found that the IADPSG diagnostic criteria would be cost-effective if associated intervention decreased the absolute incidence of preeclampsia by >0.55% and cesarean delivery by >2.7% (53). In contrast, UK health economic data show that routinely identifying GDM is not cost-effective based on perinatal outcomes (54) and that the universal WHO (IADPSG) testing approach is less cost-effective than the NICE selective screening approach (55).

Contemporary Clinical Evidence Following the Revised IADPSG GDM Diagnostic Criteria

The lack of randomized controlled trials (RCTs) evaluating outcomes in women diagnosed with GDM based on the IADPSG criteria and the clinical relevance of treating the resulting milder degrees of hyperglycemia remain controversial (56). Several retrospective studies have shown that women diagnosed with GDM by the IADPSG criteria but who were previously classified as having normal glucose tolerance were still at increased risk for obstetric and neonatal complications, including gestational hypertension, preeclampsia, cesarean delivery, macrosomia, large-for-gestational-age (LGA), shoulder dystocia, and neonatal intensive care admission, compared to women with normal glucose tolerance (57-59). For example, a 2015 retrospective study in Taiwan comparing pregnancy outcomes in women diagnosed and treated for GDM using the 2-step (GCT followed by the 100-g 3-hour OGTT) approach compared to the IADPSG 1-step approach found that the latter was associated with a reduction in gestational weight gain (GWG), birthweight, macrosomia, and LGA (60). Another retrospective study in the United Kingdom reported that women who were diagnosed with GDM based on modified IADPSG diagnostic glucose thresholds but who screened negative for GDM on 2015 NICE diagnostic criteria had a higher risk of LGA, cesarean delivery, and polyhydramnios (61). Other retrospective studies have also demonstrated higher birthweight, birthweight z-score, ponderal index, and increased rates of LGA and cesarean delivery in untreated women diagnosed with GDM based on the IADPSG criteria, compared to women with normal glucose tolerance (62,63).

The recent randomized ScreenR2GDM trial compared 1-step screening (75-g 2-hour OGTT) with 2-step screening (2 GCT thresholds ≥7.2 mmol/L and ≥7.8 mmol/L used, followed by the 100-g 3-hour OGTT) in 23 792 pregnant women in the United States (64). Despite doubling the diagnosis of GDM with the 1-step approach (16.5% vs 8.5%), there were no differences in pregnancy complications including LGA [relative risk (RR) 0.95; 97.5% CI 0.87-1.05], perinatal composite outcome (RR 1.04; 97.5% CI 0.88-1.23), gestational hypertension or preeclampsia (RR 1.00; 97.5% CI 0.93-1.08), and primary cesarean section (RR 0.98; 97.5% CI 0.93-1.02) between the different screening approaches. These findings have not resolved the diagnostic debate for GDM, with some arguing that the 1-step approach therefore demonstrates insufficient perinatal benefit for the associated increased healthcare costs (65), while others have identified potential limitations in study methodology (7,47,65,66). Despite randomization to either testing strategy, the pragmatic trial design allowed clinicians to select a preferred strategy. Consequently, one third of women randomized to the 1-step approach did not adhere to the assigned screening and were tested via the 2-step approach, compared to only 8% of women randomized to the 2-step approach. Although the study attempted to adjust for this difference using inverse probability weighting, residual provider bias cannot be excluded (47). Given this was a population level analysis of GDM screening, GDM (treatment) status differed only for the 8% of women not diagnosed with GDM based on the 2-step approach who may have otherwise been diagnosed with GDM based on the 1-step approach. Whether these women had potentially worse outcomes that may have been mitigated by treatment cannot be determined by this study. However, given the rates of pharmacotherapy were similar between the 1- and 2-step cohorts at 43% and 46%, respectively (64), this strategy detected women with essentially an equivalent risk of hyperglycemia warranting pharmacotherapy (47). This observation is consistent with other studies in UK cohorts comparing the IADPSG testing approach to the less sensitive NICE and Canadian criteria, whereby women demonstrated insulin resistance and required pharmacotherapy for control of hyperglycemia even at the most sensitive thresholds of the IADPSG diagnostic criteria (67).

More generally, the GCT fails to detect approximately 20% to 25% of women with GDM, particularly those diagnosed with GDM based on an elevated fasting glucose (68). The frequency of GDM diagnosed by the OGTT fasting glucose threshold in the HAPO study ranged from 24% to 26% in Thailand and Hong Kong to >70% in the United States (69). This highlights the variability and thus limitations of post-glucose load screening based on ethnicity. Moreover, a recent systematic review and meta-analysis of 25 studies (n = 4466 women) showed that even 1 abnormal value on the diagnostic 3-hour 100-g OGTT is associated with an increased risk of perinatal complications compared to women with a normal GCT, and this risk was similar to that of women actually diagnosed with GDM (70).

The degree of benefit of treating women with GDM defined by the IADPSG diagnostic criteria is yet to be determined. The potential benefit is inferred from the treatment of maternal hyperglycemia described in the ACHOIS and MFMU intervention trials (28,29), whereby maternal glucose levels overlapped with the thresholds recommended by the IADPSG. It is worth noting that there are differences in these 2 trials with regards to the diagnostic criteria used to define GDM and cohort characteristics (eg, women were excluded from the MFMU trial if they had an abnormal glucose screening test prior to 24 weeks’ gestation or previous GDM), and thus the generalizability of these findings in women diagnosed with GDM based on the IADPSG criteria remains contentious.

Current Classification of Hyperglycemia in Pregnancy and GDM

The WHO first defined GDM in 1965 as “hyperglycemia of diabetic levels occurring during pregnancy” (8). Thus, historically, the term “GDM” encompassed the entire spectrum of maternal hyperglycemia in pregnancy, from pregestational diabetes to hyperglycemia first detected in pregnancy. In 1979, the NDDG defined GDM as “glucose intolerance that has its onset or recognition during pregnancy” (13). This was subsequently modified in 1985 at the Second International Workshop-Conference on Gestational Diabetes as “carbohydrate intolerance resulting in hyperglycemia of variable severity with onset or first recognition during pregnancy” and remained the most widely used definition of GDM until recently (71).

Contemporary nomenclature and diagnostic criteria now more clearly differentiate between women with pregestational diabetes and those with hyperglycemia first detected in pregnancy (30) (Fig. 1). Pregestational diabetes includes type 1 diabetes, type 2 diabetes, and other types of diabetes such as cystic fibrosis-related diabetes, steroid/medication-induced diabetes, and monogenic diabetes.

Open in a separate window

Figure 1.

Flowchart summarizing the contemporary nomenclature for hyperglycemia in pregnancy.

Hyperglycemia in pregnancy is now subclassified by the IADPSG into 2 separate categories, namely “overt diabetes mellitus during pregnancy” (overt diabetes) and GDM (30). Similarly, the WHO has a binary definition of hyperglycemia in pregnancy but has replaced the term “overt diabetes” with “diabetes mellitus in pregnancy” (DIP) (11). The rationale for the IADPSG recommendation for early testing in high-risk women is to diagnose DIP early in pregnancy. This is because DIP, diagnosed based on nonpregnant diabetes glucose thresholds, recognizes the increasing prevalence of undiagnosed preexisting diabetes in women of childbearing age as well as the greater risk associated with this degree of hyperglycemia (72-74). For example, a recent study in almost 5000 women in France found that DIP was associated with a 3.5-fold greater risk of hypertensive disorders in pregnancy compared to women with normal glucose tolerance, while early‐diagnosed DIP was associated with an increased risk of congenital malformation (7.7% vs 1.0% for women with normal glucose tolerance), suggesting that early hyperglycemia in pregnancy may sometimes be present at conception (75). However, DIP is not synonymous with preexisting diabetes. In Australian, women with DIP who performed an OGTT at 6 to 8 weeks postpartum, 21% had diabetes, 38% had impaired fasting glucose or impaired glucose tolerance, and 41% returned to normal glucose tolerance (76).

Regardless of the specific nomenclature used, DIP is distinct from GDM, which is defined by lower glucose thresholds on the OGTT and was historically considered to be a condition of mid to late pregnancy. The ADA has not accepted this nomenclature and defines GDM based on timing of diagnosis: women diagnosed with diabetes in the first trimester are classified as having (preexisting) type 2 diabetes, while GDM is defined as diabetes diagnosed in later pregnancy and not meeting the diagnostic criteria for type 2 diabetes (18). A summary of the current international nomenclature and diagnostic criteria for hyperglycemia in pregnancy is presented in Table 2.

Early GDM

Most international guidelines now recommend early antenatal testing for women at high risk to identify women with DIP (11,18,30,38,39,42-44). This has resulted in increased detection of milder degrees of hyperglycemia below the threshold of DIP, referred to as GDM diagnosed prior to 24 weeks’ gestation or early GDM. Studies in women with GDM have reported that between 27% and 66% of GDM can be detected in early pregnancy depending on the population as well as the screening and diagnostic criteria used (77-81).

Recent studies evaluating the relationship between maternal glycemia and fetal growth trajectories confirm the early impact of maternal glycemia on excess fetal growth and adiposity prior to the diagnosis of standard GDM from 24 weeks’ gestation. A US multiethnic prospective cohort study of 2458 women enrolled between 8 and 13 weeks’ gestation included 107 (4.4%) women with GDM (82). GDM was associated with an increase in estimated fetal weight from 20 weeks’ gestation, which became significant at 28 weeks’ gestation. Similarly, Sovio et al showed that excessive fetal growth occurred between 20 to 28 weeks’ gestation, prior to the diagnosis of GDM, especially among women with higher body mass index [BMI (kg/m²)] (83). An Indian study also showed that excess subcutaneous abdominal adiposity was first detected at 20 weeks’ gestation, at least 4 weeks prior to the diagnosis of GDM (84). Early excess adiposity persisted despite adjustments for maternal age, BMI, GWG, fetal sex, and gestational age and remained higher at 32 weeks’ gestation (84).

Currently, there is no consensus for the preferred testing approach or diagnostic glycemic thresholds for early GDM. The IADPSG recommends diagnosing early GDM based on a fasting glucose of 5.1 mmol/L to 6.9 mmol/L (92-124 mg/dL) (30), consistent with the diagnostic fasting glucose threshold for standard GDM. The utility of a single fasting glucose measurement for early GDM diagnosis warrants consideration. First, preanalytical glucose handling variation, particularly in the setting of a single glucose measurement, is a major issue for GDM diagnostic accuracy (discussed in the following text). Second, an Israeli cohort study of 6129 women who underwent a fasting glucose test at a median of 9.5 weeks’ gestation demonstrated a positive association between first trimester fasting glucose up to 5.8 mmol/L (104.5 mg/dL) and increased risk for subsequent diagnosis of GDM, LGA, macrosomia, and cesarean section (85). Similar to the HAPO study, a clear glucose threshold was lacking, with pregnancy complications evident at fasting glucose levels <5.1 mmol/L (92 mg/dL). Third, maternal fasting glucose decreases in the first trimester, most pronounced between 6 to 10 weeks’ gestation [median decrease in glucose 0.11 mmol/L (1.98 mg/dL)] (86), while studies have consistently shown that early fasting glucose is poorly predictive of GDM at 24 to 28 weeks’ gestation (86-88), leading to potential overdiagnosis of GDM. In China, an early fasting glucose between 6.1 mmol/L to 6.9 mmol/L (110-124 mg/dL) best corresponded to later GDM diagnosis (88), but this requires further validation.

The WHO recommends the same diagnostic OGTT glucose thresholds for GDM in early pregnancy as those derived from HAPO by the IADPSG (11). However, the prognostic value of these glucose levels in early pregnancy is yet to be established. Others have proposed an hemoglobin A1c (HbA1c) risk threshold (89), based primarily on evidence that an early HbA1c ≥ 5.9% (41 mmol/mol) detected all cases of DIP and predicted adverse pregnancy outcomes in a New Zealand cohort (90). However, studies in other cohorts have found that while an elevated HbA1c in early pregnancy is highly specific, it lacks sensitivity for identifying hyperglycemia and certain perinatal complications (91,92), with no clear benefit of treating women with HbA1c 5.7% to 6.4% (39-46 mmol/mol) in early pregnancy (93,94). A summary of the various international criteria for testing of GDM in early pregnancy is presented in Table 3.

Despite the lack of diagnostic clarity for early GDM, increasing evidence suggests that women with early GDM represent a high-risk cohort (81). Early studies also reported worse pregnancy outcomes and increased insulin resistance in early GDM (78,95-97) but were confounded by the inclusion of women with pregestational diabetes. The first large retrospective cohort study excluding women with DIP showed that women diagnosed and treated for early GDM, especially those diagnosed in the first trimester, were more insulin resistant and at significantly greater risk for obstetric and neonatal complications compared to women diagnosed and treated for GDM from 24 weeks’ gestation (81). Other studies have since confirmed these findings (98,99). Concerningly, an increased risk of perinatal mortality and congenital abnormalities has also been reported in the offspring of women with early GDM (75,78,95,96), with some data demonstrating that 5% of women with early GDM have abnormal fetal echocardiograms (97). A recent meta-analysis of 13 cohort studies showed greater perinatal mortality among women with early GDM (RR 3.58; 95% CI 1.91-6.71) compared to women with a later diagnosis of GDM despite treatment (100).

A recent study assessing the pathophysiological characteristics of women diagnosed with GDM at a median of 16 weeks’ gestation compared to those diagnosed from 24 weeks’ gestation using IADPSG diagnostic criteria reported that women with early GDM had lower insulin sensitivity (defined by insulin-mediated glucose clearance during an OGTT), even after accounting for maternal BMI (101). Consistent with the pathophysiology of GDM, women with both early and standard GDM demonstrated impairment in pancreatic β-cell function (102). These data underscore GDM phenotypic differences, specifically based on timing of diagnosis and degree of hyperglycemia (103).

A key issue is the current lack of high-quality evidence that diagnosing and treating early GDM improves pregnancy outcomes. A recent major RCT in the United States evaluating early testing for GDM in 962 women with obesity included a subgroup analysis of women diagnosed and treated for GDM [early n = 69 (15.0%) vs standard n = 56 (12.1%)] based on the 2-step testing approach (104). The average gestational age at GDM diagnosis was similar at 24.3 ± 5.2 weeks for the early screen group compared to 27.1 ± 1.7 weeks in the routine screen group. There was no difference in pregnancy outcomes, although the primary composite perinatal outcome (macrosomia, primary cesarean delivery, gestational hypertension, preeclampsia, hyperbilirubinemia, shoulder dystocia, and neonatal hypoglycemia) was nonsignificantly higher in the early-screen group (56.9% vs 50.8%; P = 0.06). Requirement for insulin therapy was almost 4-fold higher, while gestational age at delivery was lower (36.7 vs 38.7 weeks’ gestation; P = 0.001) in women with early GDM. In a post hoc analysis of the Lifestyle in Pregnancy study (105), no difference in pregnancy outcomes was shown between women randomized to either lifestyle intervention (n = 36) or standard treatment (n = 54) in early pregnancy. Whether different glycemic targets are required reflecting physiological differences in early maternal glucose or whether additional risk factors contributing to a more insulin resistant phenotype such as maternal adiposity might also have a role remain unanswered (81). The ongoing Treatment of Booking Gestational Diabetes Mellitus study, evaluating the impact of immediate vs delayed care for gestational diabetes diagnosed at booking, will seek to determine whether or not there is benefit from treating early GDM (106).

The Impact of Preanalytical Glucose Processing Standards on the Diagnosis of GDM

Although the contemporary testing approach to GDM remains contentious, it is important to recognize that the diagnosis of GDM is based on the laboratory measurement of maternal glucose rather than a clinical diagnosis. Arguably then, a major issue in the contemporary diagnosis of GDM is optimizing preanalytical processing and measurement of maternal plasma glucose to ensure diagnostic accuracy (107,108). This includes optimization of sample handling and minimization of any analytic error. Unfortunately, stringent preanalytical processing standards are not currently routinely applied. The American Association for Clinical Chemistry (AACC) and ADA recommendations on laboratory testing in diabetes advise collection of plasma glucose in sodium fluoride tubes, with immediate placement in an ice slurry and centrifugation within 30 minutes (109). Citrate tubes are recommended as an alternative where early centrifugation is not possible. These standards are important because a major source of preanalytical glucose measurement error in sodium fluoride tubes is glycolysis by erythrocytes and leukocytes, which at room temperature lowers glucose levels prior to centrifugation at a rate of 5% to 7% per hour [~0.6 mmol/L (10 mg/dL)] (109,110). By 1 hour, this degree of glucose lowering is higher than the total analytical error threshold for glucose based on biological variation (107).

Recent studies have shown that OGTT preanalytical glucose processing variability greatly impacts the prevalence of GDM (67,111). Implementation of the AACC/ADA recommendations in a UK cohort resulted in higher mean glucose concentrations and 2.7-fold increased detection of GDM based on IADPSG criteria compared with the standard practice of storing sodium fluoride tubes at room temperature and delaying centrifugation until collection of all 3 OGTT samples (112). This increase in GDM diagnosis was entirely attributable to control of glycolysis (107). Similarly, in a large Australian multiethnic cohort (n = 12317), the rate of GDM diagnosis based on IADPSG criteria increased from 11.6% to 20.6% with early (within 10 minutes) vs delayed centrifugation (111). Mean glucose concentrations for the fasting, 1-hour, and 2-hour OGTT samples were 0.24 mmol/L (5.4%), 0.34 mmol/L (4.9%), and 0.16 mmol/L (2.3%) higher with early centrifugation, with the increase in GDM diagnosis primarily due to the resulting increase in fasting glucose levels (111). Importantly, the HAPO study, upon which the IADPSG diagnostic criteria for GDM was based, followed these AACC/ADA preanalytical glucose processing standards (111).

Incidence and Prevalence of GDM

GDM is 1 of the most common medical complications of pregnancy (73). In 2019, the International Diabetes Federation (IDF) estimated that 1 in 6 live births worldwide were complicated by GDM (113). More than 90% of cases of hyperglycemia in pregnancy occur in low- and middle-income countries (114), where the prevalence and severity of maternal and neonatal complications associated with GDM (47,113) contrast with the near-normal pregnancy outcomes of modern management of GDM in developed countries (115).

The prevalence of GDM varies widely, depending on the population, the specific screening and the diagnostic criteria utilized. A 2012 systematic review of the diagnostic criteria used to define GDM reported a worldwide prevalence of GDM of 2% to 24.5% for the WHO criteria, 3.6% to 38% for the Carpenter and Coustan criteria, 1.4 to 50% for the NDDG criteria, and 2% to 19% for the IADPSG criteria (116).

Regardless of the specific diagnostic criteria or population, the prevalence of GDM continues to rise internationally, corresponding to epidemiological factors including the background rates of type 2 diabetes and increased incidence of obesity in women of childbearing age and rising maternal age (117-124). Implementation of the revised IADPSG diagnostic criteria have further increased the proportion of women being diagnosed with GDM (69,125,126). The incidence of GDM in the original HAPO study cohort applying the IADPSG diagnostic criteria ranged from 9.3% to 25.5% depending on study site (69). Recent international prevalence data also demonstrate marked variability in the rate of GDM, ranging from 6.6% in Japan and Nepal to 45.3% of pregnancies in the United Arab Emirates (127).

Risk Factors for GDM

Several modifiable and nonmodifiable risk factors for GDM have been identified (Table 4). A history of GDM in a previous pregnancy is the strongest risk factor for GDM, with reported recurrence rates of up to 84% (128). The risk of recurrence varies greatly depending on ethnicity (128). Ethnicities at increased risk for development of type 2 diabetes, such as South and East Asians, Hispanic, Black and Native Americans, Aboriginal and Torres Strait Islanders, and Middle Easterners are also associated with an increased risk of GDM (129,130). A US study of over 123 000 women reported the prevalence of GDM using the 2000 ADA diagnostic criteria to be the highest among Filipinas (10.9%) and Asians (10.2%), followed by Hispanics (6.8%), non-Hispanic Whites (4.5%) and Black Americans (4.4%) (131). Women who have had GDM are at increased risk for subsequent type 2 diabetes, while family history of type 2 diabetes in a first-degree relative or sibling with GDM is a major risk factor for GDM (129,132-134).

Increasing maternal age is also a risk factor for GDM (129,133-135). The prospective First and Second Trimester Evaluation of Risk trial (n = 36 056) demonstrated a continuous positive relationship between increasing maternal age and risk for adverse pregnancy outcomes, including GDM (135). Maternal age 35 to 39 years and ≥40 years was associated with an adjusted odds ratio (OR) for GDM of 1.8 (95% CI 1.5-2.1) and 2.4 (95% CI 1.9-3.1), respectively (135). Other studies in high-risk cohorts have reported a lesser risk between increasing maternal age and GDM after adjustment for other risk factors (136).

Maternal prepregnancy overweight (BMI 25-29.99 kg/m²) or obesity (BMI ≥ 30 kg/m²) are common risk factors for GDM (129,130,133,134,136,137). The risk of GDM is increased almost 3-fold (95% CI 2.1-3.4) in women with class I obesity (BMI 30-34.99 kg/m²) and 4-fold (95% CI 3.1-5.2) in women with class II obesity (BMI 35-39.99 kg/m²), compared to women with a BMI < 30 kg/m² (138). High GWG, particularly in the first trimester, is also associated with an increased risk for GDM (131,139,140). Further, women with obesity and high GWG are 3- to 4-fold more likely to develop abnormal glucose tolerance compared to women who remained within the 1990 Institute of Medicine (IOM) recommendations for GWG (131,141). Interpregnancy weight gain is also a risk factor for GDM and perinatal complications in a subsequent pregnancy (142) and may be a potential confounder when considering the risk of GDM recurrence.

Studies have demonstrated an association between polycystic ovary syndrome and GDM, although this is significantly attenuated after adjustment for maternal BMI (143,144). Other risk factors for GDM include multiparity (133,134), twin pregnancy (145,146), previous macrosomia (123), a history of perinatal complications (134), maternal small-for-gestational-age (SGA) or LGA (134), physical inactivity (129,147,148), low-fiber high-glycemic load diets (149), greater dietary fat and lower carbohydrate intake (137), and medications such as glucocorticoids and anti-psychotic agents (150,151). Maternal pre- and early pregnancy hypertension is also associated with an increased risk of developing GDM (152,153).

Overall, noting the variation in performance and utility of clinical risk factors based on local population factors, previous GDM and family history of diabetes appear to be the strongest clinical risk factors for GDM (154-157). Ethnicity, higher maternal age, and BMI are also strong predictors for GDM (154-158).

Pathophysiology of GDM

Normal pregnancy is associated with marked changes in glycemic physiology (159,160). There is a progressive increase in insulin resistance, predominantly due to increased circulating placental hormones including growth hormone, corticotrophin-releasing hormone, human placental lactogen, prolactin, estrogen, and progesterone (161-166). Increased maternal adiposity particularly in early pregnancy also promotes insulin resistance, contributing to facilitated lipolysis by late pregnancy (167,168). The resultant increase in maternal free fatty acid (FFA) levels exacerbates maternal insulin resistance by inhibiting maternal glucose uptake and stimulating hepatic gluconeogenesis (168,169). By late pregnancy, studies have reported decreases in maternal glucose sensitivity between 40% and 80% in women with normal or increased BMI (170-172). Increased maternal insulin resistance results in higher maternal postprandial glucose levels and FFAs for maternal growth (164,167,173) and increased facilitated diffusion across the placenta, leading to greater availability of glucose for fetal growth (161,174). This progressive rise in maternal insulin resistance underpins the delayed testing approach to GDM, aiming to maximize detection of GDM when insulin resistance is at its greatest in mid- to late gestation.

In addition to increased insulin resistance and elevated postprandial glucose, adaptations in normal pregnancy include enhanced insulin secretion (160,165). Maternal glucose levels are maintained at lower levels than in healthy nonpregnant women (175,176), and euglycemia is maintained by a corresponding 200% to 250% increase in insulin secretion, most notable in early pregnancy (161,167,177). Human placental lactogen, in addition to prolactin and growth hormone, primarily regulate increased maternal β-cell insulin secretion and proliferation during pregnancy (178-180). Rodent studies have demonstrated a 3- to 4-fold increase in β-cell mass during pregnancy, mediated via hypertrophy, hyperplasia, neogenesis, and/or reduced apoptosis (181,182).

GDM is characterized by a relative insulin secretory deficit (177), in which maternal β-cell insulin secretion is unable to compensate for the progressive rise in insulin resistance during pregnancy (183). This leads to decreased glucose uptake, increased hepatic gluconeogenesis, and maternal hyperglycemia (167). It is hypothesized that this results from the failure of β-cell mass expansion (182,184). Hyperlipidemia, characterized predominantly by higher serum triglycerides, may also cause lipotoxic β-cell injury, further impairing insulin secretion (185,186). The pathogenesis of GDM therefore parallels that of type 2 diabetes, characterized by both increased insulin resistance and relative insulin deficiency arising from a reduction in β-cell function and mass (187,188).

Serial studies of the insulin secretory response in women who develop GDM suggest that the abnormal insulin secretory response is present from prepregnancy and increases in early pregnancy, prior to and independent of changes in insulin sensitivity (170,189-191). These data suggest that many women with GDM may have chronic or preexisting β-cell dysfunction, potentially mediated by circulating hormones including leptin (191).

Genetics of GDM

The genetics of GDM and glucose metabolism in pregnancy remain poorly defined. Data on epigenetic mechanisms in GDM are especially lacking and primarily limited to the potential role of DNA methylation in mediating the intrauterine effects of GDM on offspring outcomes (192,193).

Most genetic studies have focused on variants associated with type 2 diabetes and have demonstrated a similar association with GDM (194,195). A meta-analysis of 28 case-control studies (n = 23425) (196) identified 6 genetic polymorphisms at loci involved in insulin secretion [insulin-like growth factor 2 messenger RNA-binding protein 2 (IGF2BP2), melatonin receptor 1B (MTNR1B) and transcription factor 7-like 2 (TCF7L2)] (197-199), insulin resistance [insulin receptor substrate 1 (IRS1) and peroxisome proliferator-activated receptor gamma (PPARG)] (200,201), and inflammation [tumor necrosis factor alpha (TNF-α)] (202) in type 2 diabetes. Overall, only MTNR1B, TCF7L2, and IRS1 were also significantly associated with GDM, supporting the role of both impaired insulin secretion and insulin resistance in the pathogenesis of GDM as well as type 2 diabetes (196). Subgroup analysis showed the risk alleles of TCF7L2 and PPARG were significant only in Asian populations, while the association between IRS1 and TCF7L2 and GDM risk varied depending on diagnostic criteria and genotype methodology (196), highlighting the need for further large confirmatory studies.

Two genome-wide association studies (GWAS) have evaluated the genetic associations for GDM and glucose metabolism (194,203). The first, a 2-stage GWAS in Korean women, compared 468 women with GDM and 1242 normoglycemic women using 2.19 million genotyped markers before further genotyping 11 loci in 1714 women, identifying 2 loci significantly associated with GDM (203). A variant in cyclin-dependent kinase 5 regulatory subunit-associated protein 1-like 1 (CDKAL1) had the strongest association with GDM, followed by a variant near MTNR1B expressed in pancreatic β-cells (204). The IGF2BP2 variant did not reach genome-wide significance with GDM in this study. CDKAL1 was significantly associated with decreased fasting insulin concentration and homeostasis model assessment of β-cell function in women with GDM, consistent with impaired β-cell compensation. MTNR1B was associated with decreased fasting insulin concentrations in women with GDM and increased fasting glucose concentrations in both women with and without GDM (203). Variants in CDKAL1 and MTNR1B have previously been associated with type 2 diabetes risk (205,206).

A subsequent GWAS performed in a subset of the HAPO cohort (n = 4528) comprising European, Thai, Afro-Caribbean, and Hispanic women evaluated maternal metabolic traits in pregnancy (194). This study reported 5 variants associated with quantitative glycemic traits in the general population (207,208) that were also associated with glucose or C-peptide levels in pregnancy, although strength of association varied across cohorts (194). Specifically, loci in glucokinase regulator (GCKR), glucose-6-phosphatase 2 (G6PC2), proprotein convertase subtilisin/kexin type 1 (PCSK1), protein phosphatase 1, regulatory subunit 3B (PPP1R3B), and MTNR1B were associated with fasting glucose. In addition, GCKR and PPP1R3B were associated with fasting C-peptide levels, while MTNR1B was associated with 1-hour postload glucose. These loci have also previously been associated with lipid metabolism (GCKR and PPP1R3B), glycogen metabolism (PPP1R3B), and obesity-related traits (PCSK1) (209-214).

Two additional novel loci identified near hexokinase domain containing 1 (HKDC1) associated with 2-hour postload glucose, and β-site amyloid polypeptide cleaving enzyme 2 (BACE2) associated with fasting C-peptide, demonstrated limited association with glycemic traits outside of compared to in pregnancy (215). In general, however, studies evaluating associations between genetic risk scores, glycemic traits in pregnancy, and GDM have also confirmed that genetic determinants of fasting glucose and insulin, insulin secretion, and insulin sensitivity reported outside of pregnancy influence GDM risk (216). A summary of the genes associated with GDM is provided in Table 5.

Maturity-onset Diabetes of the Young

Maturity-onset diabetes of the young (MODY) is the most common form of monogenic diabetes; inherited forms of diabetes characterized by defects in single genes regulating β-cell development and function (217,218). MODY consists of several autosomal dominant forms of diabetes accounting for up to 2% of all diabetes diagnoses (219). A diagnosis of MODY requires confirmatory molecular genetic testing, and thus MODY is frequently misdiagnosed as preexisting diabetes or GDM, accounting for up to 5% of GDM “cases” (220-223). A UK study reported that HNF-1α (MODY3) (52%) and glucokinase (GCK)-MODY subtype (MODY2) (32%) were most frequent in probands confirmed with MODY, followed by HNF-4α (MODY1) and HNF-1β (MODY5) (224).

Women with GCK-MODY often first present following antenatal screening for GDM, with an estimated prevalence of 1% of all GDM “cases” actually GCK-MODY (220,222). GCK-MODY is caused by mutations in the glucokinase gene, leading to a greater set point for glucose stimulated insulin release (219). Clinically, GCK-MODY is defined by mild, stable fasting hyperglycemia [fasting glucose 98-150 mg/dL (5.4-8.3 mmol/L)] and low rates of microvascular and macrovascular complications (220). It should be suspected following a positive OGTT in pregnancy if the fasting glucose is ≥5.5 mmol/L, the glucose increment from the fasting to 2-hour (75-g) OGTT is small (<4.6 mmol/L), and there is a positive family history of mild hyperglycemia or diabetes. In addition, a combination of fasting glucose ≥ 100 mg/dL (5.6 mmol/L) and BMI < 25 kg/m² has been shown to have a sensitivity of 68% and a specificity of 99% for differentiating GCK-MODY from GDM (220). Importantly, management differs from that of GDM because the need for intensive maternal glycemic control largely depends on whether the GCK-MODY mutation is also present in the fetus (220,225,226). Maternal insulin therapy is therefore only recommended in the presence of increased fetal abdominal growth (>75th centile) measured on serial ultrasounds from 26 weeks’ gestation, as this indicates that the fetus does not have the GCK mutation (220).

Consequences of GDM

GDM is associated with excess neonatal and maternal short- and long-term morbidity, summarized in Table 6.

Neonatal Complications

The Pedersen hypothesis describes the pathophysiology contributing to perinatal complications in GDM (229). Maternal hyperglycemia results in fetal hyperglycemia via facilitated diffusion of glucose by the glucose transporter 1 (GLUT1) (230). Fetal hyperglycemia results in fetal hyperinsulinemia, promoting fetal anabolism, excessive fetal adiposity, and accelerated growth, leading to LGA and macrosomia (231-239). Maternal hyperlipidemia also contributes to excess fetal growth (233,240). Macrosomia and LGA increase the risk of cesarean section, birth trauma, and perinatal complications including shoulder dystocia, brachial plexus injury and fracture, and perinatal asphyxia (27,132,237,238,241-243). Increased risk of perinatal asphyxia is associated with fetal death in utero, polycythemia, and hyperbilirubinemia (27,244-246). Fetal hyperinsulinemia can also increase the risk of metabolic abnormalities including neonatal hypoglycemia, hyperbilirubinemia, and respiratory distress syndrome postpartum (27,244). The risk appears to be greater among offspring of women with more severe hyperglycemia (247). Figure 2 summarizes the perinatal consequences of GDM.

Open in a separate window

Figure 2.

Perinatal consequences of gestational diabetes mellitus.

Short-term Risk

In the HAPO study, higher maternal glucose levels were associated with an increased risk of LGA, shoulder dystocia or birth injury, and neonatal hypoglycemia (27). A recent systematic review (n = 207 172) confirmed similar positive linear associations for maternal glycemia based on maternal glucose thresholds for the GCT, 75-g 2-hour OGTT, or 100-g 3-hour OGTT and risk of cesarean section, induction of labor (IOL), LGA, macrosomia, and shoulder dystocia (248). GDM has also been associated with an increased risk of preterm birth, birth trauma, neonatal respiratory distress syndrome, and hypertrophic cardiomyopathy (27,244,249). An increased risk of congenital malformations in the offspring has been reported, although whether this persists after adjustment for maternal age, BMI, ethnicity, and other contributing factors is unknown (250). A French cohort study (n = 796 346) reported a 30% higher risk of cardiac malformations in the offspring of women with GDM compared to women with normal glucose tolerance, after excluding women with likely undiagnosed pregestational diabetes (249). However, this increased risk only reached statistical significance in women treated with insulin therapy. Maternal BMI, which was not evaluated in these studies, may account for these findings (251,252). Similarly, a reported increase in perinatal mortality after 35 weeks’ gestation in the offspring of women with GDM may also be confounded by obesity (253-256). An increased risk of perinatal mortality after 37 weeks’ gestation was demonstrated in French women with GDM on dietary intervention, possibly because these women delivered later than women treated with insulin therapy (249). In contrast, the HAPO study did not demonstrate excess perinatal mortality in their untreated cohort (27).

Modern management of GDM and associated maternal risk factors is associated with near-normal birthweight in developed countries (115,257). This is important because birthweight is the major risk factor for shoulder dystocia, brachial plexus injury, neonatal hypoglycemia, and neonatal respiratory distress syndrome in the offspring of women with and without GDM (242). A retrospective cohort study of 36 241 pregnancies in the United States reported that the risk of shoulder dystocia among infants of women without GDM compared to women with GDM was 0.9% vs 1.6% if birthweight was <4000 g and 6.0% vs 10.5% if birthweight was ≥4000 g (macrosomia) (242). The risk of neonatal hypoglycemia in infants with birthweight < 4000 g was 1.2% vs 2.6% and 2.4% vs 5.3% for birthweight ≥ 4000 g, in women without GDM compared to women with GDM, respectively. Similar findings were seen for brachial plexus injury and neonatal respiratory distress syndrome. Thus, GDM confers increased risk of perinatal complications independent of birthweight.

The risk of stillbirth is also greater in women with GDM. A large US retrospective analysis examined stillbirth rates at various stages of gestation in over 4 million women, including 193 028 women with GDM. The overall risk of stillbirth from 36 to 42 weeks’ gestation was higher in women with GDM compared to women without GDM (17.1 vs 12.7 per 10 000 deliveries; RR 1.34; 95% CI 1.2-1.5) (253). This increased risk of stillbirth was also observed at each gestational week: 3.3 to 8.6 per 10 000 ongoing pregnancies in women with GDM compared to 2.1 to 6.4 per 10 000 ongoing pregnancies in women without GDM from 36 to 41 weeks’ gestation (253). For women with GDM, the relative risk of stillbirth was highest in week 37 (RR 1.84, 95% CI 1.5-2.3). Notably, the risk of stillbirth is highest in women with undiagnosed GDM. In a UK prospective case-control study (n = 1024), women with undiagnosed GDM based on a fasting glucose level ≥ 5.6mmol/L (≥100 mg/dL) had a 4-fold greater risk of late stillbirth (defined as occurring ≥28 weeks’ gestation) compared to women with fasting glucose < 5.6mmol/L (<100 mg/dL) (74). In contrast, women at risk of GDM based on NICE risk factors who were diagnosed with GDM on the OGTT had a similar risk of stillbirth to women who were not at risk of GDM. This suggests that diagnosing and managing GDM reduces the risk of stillbirth to near-normal levels (74).

Long-term Risk in the Offspring

Recent epidemiological studies suggest an increased risk of later adverse cardiometabolic sequelae in the offspring of women with GDM (227,258). A large Danish population-based cohort study (n = 2 432 000) demonstrated an association between maternal diabetes and an increased rate of early onset cardiovascular disease (CVD; ≤40 years of age) among offspring (259). GDM specifically was associated with a 19% increased risk of early onset CVD (95% CI 1.07-1.32). A longitudinal UK study provides potential mechanistic insight, finding that GDM was associated with alterations in fetal cardiac function and structure, with reduced systolic and diastolic ventricular function persisting in infancy (260). This is consistent with the association between in utero exposure to maternal hyperglycemia and fetal programming first reported in the Native American Pima population, characterized by a high prevalence of obesity, type 2 diabetes, and GDM (261).

The recent HAPO Follow Up Study (HAPO-FUS), which was not confounded by treatment of maternal glycemia, included 4832 children 10 to 14 years of age whose mothers were participants of HAPO (227). The HAPO-FUS demonstrated a durable impact of maternal glycemia with long-term offspring glucose metabolism, including at glucose levels lower than those diagnostic for GDM (227). A generally linear relationship between maternal antenatal glucose and offspring glucose levels and related outcomes was observed. Increasing maternal glucose categories were associated with a higher risk of impaired fasting glucose and impaired glucose tolerance and higher timed glucose measures and HbA1c levels and were inversely associated with insulin sensitivity and disposition index by 14 years of age, independent of maternal and childhood BMI and family history of diabetes (227). A positive association was observed between GDM defined by any criteria and glucose levels and impaired glucose tolerance in the offspring at ages 10 to 14 years and an inverse association with offspring insulin sensitivity (262). Higher frequencies of childhood obesity and measures of adiposity across increasing categories of maternal OGTT glucose levels were also noted (262). Recent evidence for increased glucose-linked hypothalamic activation in offspring aged 7 to 11 years previously exposed to maternal obesity and GDM in utero, which predicted higher subsequent BMI, represents 1 possible mechanism for this increased childhood obesity risk (263).

Maternal Complications

Management of GDM

Lifestyle Intervention

The main objective of GDM management is to attain maternal normoglycemia because evidence suggests that excessive fetal growth can be attenuated by maintaining near normal glucose levels (286,287). The foundation of this approach is medical nutrition therapy. Given carbohydrates are the primary determinant of maternal postprandial glucose levels, current dietary practice aims to modify carbohydrate quality (glycemic index) and distribution (32,288,289). The original nutritional approach for GDM decreased total carbohydrate intake to 33% to 40% of total energy intake (EI) and was associated with reduced postprandial glycemia and fetal overgrowth (290). More recent evidence suggests that higher carbohydrate intake and quality (lower glycemic index) between 60% and 70% EI can also limit maternal hyperglycemia (291-293). Nevertheless, there remain limited data to support a specific dietary intervention for GDM (294). A recent meta-analysis (18 RCTs; n = 1151) showed that enhancing nutritional quality (modified dietary intervention, defined as a dietary intervention different from the usual one used in the control group) after GDM diagnosis, irrespective of the specific dietary approach, improved maternal fasting and postprandial glycemia, and reduced pharmacotherapy requirements, birthweight, and macrosomia (295).

Guidelines therefore currently recommend a range of carbohydrate intake between 33% and 55% EI (32,288,289). Studies have reported improved pregnancy outcomes in GDM with both lower carbohydrate (42%E) and high‐carbohydrate (55%E) diets (296), reflected in the most recent Academy of Nutrition and Dietetics guidelines, which state that beneficial effects on pregnancy outcomes in GDM are seen with a range of carbohydrate intakes (288). The IOM guidelines recommend a carbohydrate intake of at least 175 g/day and a total daily caloric intake of 2000 to 2500 kilocalories during pregnancy (289). The ACOG recommends a lower carbohydrate diet (33-40%E) (297). However, the ADA has raised concerns over the corresponding higher maternal fat intake, fetal lipid exposure, and overgrowth resulting from lowering carbohydrate intake (298) and withdrew specific dietary guidelines for GDM in 2005 (299).

Given maternal glucose primarily supports fetal growth and brain development (300), theoretically if the maternal diet is too low in carbohydrate, the maternal-fetal glucose gradient may be compromised. Restriction of total maternal EI is associated with reduced fetal growth (301). A recent systematic review similarly showed that lower carbohydrate intake correlated with lower birthweight and greater incidence of SGA (302), with a lower carbohydrate threshold of 47% EI associated with appropriate fetal growth (302,303). Importantly, the lower carbohydrate threshold independent of energy restriction in GDM is yet to be established. Related safety concerns with lower carbohydrate diets include the potential risk of higher fetal exposure to maternal ketones (304) and micronutrient deficiency (305,306). In vitro studies have shown that ketones suppress trophoblast uptake of glucose, jeopardizing glucose transfer across the placenta (307). Clinically, a prospective US cohort study of women with preexisting diabetes, GDM, or normal glucose tolerance demonstrated an inverse correlation between higher maternal third trimester beta-hydroxybutyrate and FFAs and lower offspring intellectual development scores at 2 to 5 years of age, although total carbohydrate, EI, and maternal BMI were not reported (304).

Gestational Weight Gain

The IOM has published recommendations for weight gain during pregnancy based on prepregnancy BMI (289), but no specific recommendations for weight gain in GDM exist (286). In women with overweight or obesity, studies have suggested that weight reduction or gain ≤ 5 kg increased the risk of SGA (308). A recent systematic review based on data from almost 740 000 women demonstrated that GWG of 5 kg to 9 kg in women with class I obesity (BMI 30-34.99 kg/m²), 1 to <5 kg for class II obesity (35-39.99 kg/m²), and no GWG for women with class III obesity (BMI ≥ 40kg/m²), minimized the combined risk of LGA, SGA, and cesarean section (309).

A meta-analysis (n = 88 599) evaluating the relationship between GWG and pregnancy outcomes in GDM specifically showed that GWG greater than the IOM recommendations was associated with an increased risk of pharmacotherapy, as well as of hypertensive disorders of pregnancy, cesarean section, LGA, and macrosomia (310). GWG below the IOM recommendations was protective for LGA (RR 0.71; 95% CI 0.56-0.90) and macrosomia (RR 0.57; 95% CI 0.40-0.83) and did not increase the risk of SGA (RR 1.40; 95% CI 0.86-2.27) (289). This suggests that GWG targets in GDM may need to be lower than the current recommendations for normal pregnancy. However, from a practical perspective, only 30% of women gained less than the recommended IOM GWG targets (310).

Maternal Glucose Targets

Fasting and postprandial glucose testing with either the 1- or 2-hour postprandial glucose value is recommended in women with GDM. The 1-hour postprandial glucose approximates to the peak glucose excursion in pregnancy in women without diabetes and those with type 1 diabetes (175). Studies have shown that the 1-hour postprandial peak glucose level correlates with amniotic fluid insulin levels, reflecting fetal hyperinsulism (311) and with fetal abdominal circumference in women with type 1 diabetes (286). An RCT that compared pre- to postprandial maternal SMBG values showed that titrating insulin therapy based on the 1-hour postprandial values was associated with improved maternal glycemic control and may better attenuate the risk of neonatal complications attributed to fetal hyperinsulinemia (312).

Treatment targets based on maternal SMBG levels vary internationally (Table 7). There is some suggestion that lower glucose targets may improve pregnancy outcomes in GDM (176,313,314), but this is yet to be evaluated in adequately powered RCTs. Conversely, lower glycemic targets may be associated with an increased risk of SGA (315-317) and maternal and fetal hypoglycemia (318,319). A small study evaluating stringent glycemic targets in 180 women with GDM failed to demonstrate additional benefits, with no differences in the rates of cesarean section, birthweight, macrosomia, or SGA in the offspring of women randomized to intensive [preprandial glucose ≤ 5.0 mmol/L (90 mg/dL) and 1-hour postprandial glucose ≤ 6.7 mmol/L (121 mg/dL)] compared to standard treatment targets [preprandial glucose ≤ 5.8 mmol/L (104.5 mg/dL) and 1-hour postprandial glucose ≤ 7.8 mmol/L (140 mg/dL)] (320).

Insulin Therapy

Insulin has traditionally been the preferred treatment for GDM if maternal glucose levels remain elevated on medical nutrition therapy (267). Depending on targets, approximately 50% of women with GDM are prescribed insulin therapy to maintain normoglycemia (321,322), with a combination of evening intermediate-acting insulin if fasting glucose levels are elevated and mealtime rapid-acting insulin when indicated. Additional daytime intermediate-acting insulin may also be needed to control prelunch or predinner hyperglycemia.

Decreasing insulin doses in the third trimester may simply reflect the physiological increase in maternal insulin sensitivity observed at this stage of pregnancy (176,323). However, substantial insulin dose reduction, recurrent maternal hypoglycemia, and/or slowing of fetal growth or preeclampsia may indicate underlying pathophysiological placental insufficiency (324), impacting the timing of delivery and intensity of obstetric monitoring.

Risk factors for insulin therapy include earlier diagnosis of GDM (81), the pattern and degree of elevation of the 75-g 2-hour OGTT diagnostic glucose thresholds (325), and ethnicity (325). Other risk factors including gestational age and HbA1c level at the time of GDM diagnosis, BMI, and family history of diabetes account for only 9% of the attributable risk for insulin therapy (321). A recent Australian study found that maternal age > 30 years, family history of diabetes, prepregnancy obesity, previous GDM, early diagnosis of GDM, fasting glucose ≥ 5.3 mmol/L (96 mg/dL) and HbA1c ≥ 5.5% (37 mmol/mol) at diagnosis were all independent predictors for insulin therapy (326). Insulin usage could also be estimated according to the number of predictors present, with up to 93% of women with 6 to 7 predictors using insulin therapy compared with less than 15% of women with 0 to 1 predictors (326).

Oral Pharmacotherapy

Oral pharmacotherapy options include glyburide and metformin. Oral pharmacotherapy is associated with improved cost effectiveness, compliance, and acceptability compared to insulin therapy (327). However, there are issues regarding efficacy and safety, particularly longer term, and thus insulin is generally preferred as first-line pharmacotherapy following lifestyle intervention.

Glyburide is commonly prescribed as first-line therapy for GDM in the United States (328). An early study evaluating the efficacy of glyburide vs insulin therapy in 404 women with GDM reported no differences in maternal glucose levels or neonatal outcomes between the treatment groups (329). However, subsequent studies show that approximately 20% of women treated with glyburide required additional insulin therapy to achieve adequate maternal glycemia (330). Moreover, a large retrospective US study of almost 111 000 women with GDM, in which 4982 women were treated with glyburide and 4191 women were treated with insulin, reported that glyburide was associated with an increased risk of neonatal complications including neonatal intensive care admission, respiratory distress syndrome, hypoglycemia, birth injury, and LGA compared to insulin therapy (331). Although transplacental transfer of glyburide to the fetus is highly variable, it can reach 50% to 70% of maternal plasma concentration (332), potentially causing direct stimulation of fetal insulin production (333).

The use of metformin in pregnancy continues to rise (334). However, its use remains controversial, due to the potential concerns regarding long-term metabolic programming effects of placental transfer of metformin to the fetus, with some studies suggesting similar plasma concentrations of metformin in the maternal and fetal circulation (335). A recent systematic review and meta-analysis of 28 studies (n = 3976) evaluating growth in offspring of women with GDM exposed to metformin compared to insulin therapy found that neonates exposed to metformin had lower birthweights (mean difference −107.7 g; 95% CI −182.3 to −32.7), decreased risk of LGA (OR 0.78; 95% CI 0.62-0.99), and macrosomia (OR 0.59; 95% CI 0.46-0.77) and lower ponderal indices than neonates whose mothers were treated with insulin (336). No difference in the risk of SGA was found, in contrast to outcomes in women with type 2 diabetes, with the Metformin in Women with Type 2 Diabetes RCT observing more than double the rate of SGA (95% CI 1.16-3.71) in the metformin treated cohort, in association with lower insulin doses, HbA1c, and GWG (337). Offspring of women with GDM exposed to metformin also demonstrate accelerated postnatal growth at 18 to 24 months of age (2 studies; n = 411; mean difference in weight 440 g; 95% CI 50-830), resulting in higher BMI at 5 to 9 years of age (3 studies; n = 520; BMI mean difference 0.78 kg/m², 95% CI 0.23-1.33) (336).

The Metformin in Gestational Diabetes trial randomized 751 women to receive either metformin or insulin therapy, finding no significant difference in the composite neonatal outcome of neonatal hypoglycemia, respiratory distress syndrome, hyperbilirubinemia, low Apgar scores, birth trauma, and preterm birth (322). There was a trend toward increased preterm birth and decreased maternal GWG in women treated with metformin, while severe neonatal hypoglycemia was highest in those treated with insulin. Almost 50% of women treated with metformin required the addition of insulin therapy (322). Other studies have reported that between 14.0% and 55.8% of women treated with metformin also require insulin therapy to achieve optimal glycemic control (338,339). The Metformin in Gestational Diabetes: The Offspring Follow-Up 2-year follow-up study found that children exposed to metformin had increased subcutaneous fat localized to the arm compared with children whose mothers were treated with insulin alone (340). By 7 and 9 years of age the children exposed to metformin had similar offspring total and abdominal body fat percentage and metabolic biochemistry including fasting glucose, insulin, and lipids but were larger overall based on measures including weight, arm and waist circumference, waist-to-height ratio, and dual-energy X-ray absorptiometry fat mass and lean mass (341). These findings are consistent with a recent follow-up study of metformin therapy in pregnant women with polycystic ovary syndrome, which showed that children exposed to metformin in utero had higher BMI and rates of overweight and obesity at 4 years of age (342).

A recent Cochrane review (8 RCTs; n = 1487) evaluating the use of metformin, glyburide, and acarbose in women with GDM found that the benefits and potential harms of these therapies in comparison to each other are unclear (343). Other meta-analyses comparing glyburide, metformin, and insulin have shown that metformin was associated with lower GWG, gestational hypertension, and postprandial maternal glucose levels compared to either glyburide or insulin (344,345), but metformin was associated with an increased risk of preterm birth compared to insulin (345). Compared to metformin, glyburide was associated with a higher risk of increased birthweight, LGA, macrosomia, neonatal hypoglycemia, and increased GWG (344). More recently, a small RCT (n = 104) suggested that glyburide and metformin were comparable in terms of maternal glycemia and perinatal outcomes (346). Treatment success after second-line (oral) therapy was higher in the (first-line) metformin vs glyburide cohort (87% vs 50%; P = 0.03), suggesting that metformin may be the preferred first-line therapy. Overall, most women required either a combination of metformin and glyburide to achieve glycemic control and/or replacement of first-line oral therapy due to hypoglycemia and gastrointestinal side effects, suggesting neither agent alone is likely to be successful in most women with GDM. Combined oral pharmacotherapy had an efficacy rate of 89%, with only 11% of women required third-line therapy with insulin (346). However, the effects of dual oral therapy crossing the placenta on long-term potential fetal programming via their effects on cellular metabolism, hepatic gluconeogenesis, and insulin sensitivity (metformin) (347) and fetal hyperinsulinemia (glyburide) is unknown (348).

Obstetric Management

A recent Cochrane review consisting of only 3 small RCTs (n = 524) reported insufficient (very low certainty) evidence to evaluate the use of fetal biometry in guiding the medical management of GDM (349). Nevertheless, serial fetal growth ultrasounds, particularly assessing fetal abdominal circumference, are potentially useful in guiding the intensity of maternal glucose targets and insulin therapy (350-352). Studies have demonstrated that neonates with an estimated fetal weight ≥ 75th percentile on early third trimester ultrasound were 10-fold more likely to be LGA compared to neonates with an estimated fetal weight < 75th percentile (353). Measured fetal abdominal circumference < 90th percentile on 2 ultrasounds at 3- to 4-week intervals has also been shown to provide high reliability in excluding the risk of LGA (351). Moreover, a recent retrospective study (n = 275) found that estimated fetal weight or abdominal circumference up to the 30th percentile on third trimester ultrasound was associated with a greater risk of adverse neonatal outcomes, comparable to that observed with abdominal circumference or estimated fetal weight > 95th percentile in women with hyperglycemia in pregnancy (including GDM) (354). These findings suggest the potential utility of fetal biometry at thresholds other than defining SGA or LGA in identifying higher risk pregnancies in GDM.

The optimal timing of delivery in GDM is complex, guided by maternal glycemic control in addition to maternal and fetal factors, and has not been definitively established. Current guidelines recommend delivery by 40 + 6 weeks’ gestation in low-risk women with GDM managed with diet alone and from 39 + 0 to 39 + 6 weeks’ gestation for women with GDM well controlled with therapy (38,277,355). A recent Canadian population-based cohort study examining the week-specific risks of severe pregnancy complications in women with diabetes included 138 917 women with GDM and 2 553 243 women without diabetes over a 10-year period (356). There was no significant difference in gestational age-specific maternal mortality or morbidity (defined as ≥1 of the following in the immediate perinatal period: obstetric embolism, obstetric shock, postpartum hemorrhage with hysterectomy or other procedures to control bleeding, sepsis, thromboembolism, or uterine rupture) between iatrogenic delivery and expectant management in women with GDM. However, iatrogenic delivery was associated with an increased risk of neonatal mortality and morbidity (birth or fetal asphyxia, grade 3 or 4 intraventricular hemorrhage, neonatal convulsions, other disturbances of cerebral status of newborn, respiratory distress syndrome, birth injury, shoulder dystocia, stillbirth or neonatal death) at 36 to 37 weeks’ gestation (76.7 and 27.8 excess cases per 1000 deliveries, respectively) but a lower risk of neonatal morbidity and mortality at 38 to 40 weeks’ gestation (7.9, 27.3, and 15.9 fewer cases per 1000 deliveries, respectively) compared with expectant management, suggesting that delivery at 38, 39, or 40 weeks’ gestation may provide the best neonatal outcomes in women with GDM (356).

Longer Term Management of Women Following GDM

Up to one third of women with GDM diagnosed by pre-IADPSG criteria will have glucose levels consistent with diabetes or prediabetes on postpartum testing at 6 to 12 weeks (357). Thus, a repeat OGTT or fasting glucose as early as 6 to 12 weeks’ postpartum is recommended to confirm maternal glucose status (41,277). Only around 25% of women are tested at this time point with compliance with postpartum testing ranging between 23% and 58% (357,358). In women with GDM with overweight or obesity, a reduction in interpregnancy BMI of ≥2.0 kg/m² reduces the risk of subsequent GDM by 74% (359). Longer term, women should perform regular cardiometabolic health assessment and optimization of lifestyle measures to reduce their greater risk of type 2 diabetes and cardiovascular disease (282,360,361). Up to 74% of women with obesity and previous GDM develop type 2 diabetes compared with <25% of women who achieve a normal BMI postpartum following GDM (362). It is unclear how relevant these studies in older women are for current clinical care given recent data that 50% of women develop type 2 diabetes within 5 to 10 years post-GDM diagnosis (273). The Diabetes Prevention Program demonstrated that lifestyle intervention and metformin therapy improved insulin sensitivity and preserved β-cell function in women with a history of previous GDM (363). Early type 2 diabetes prevention following GDM is therefore an essential component of the contemporary GDM detection and management paradigm (276).

Treatment of GDM and Long-term Offspring Outcomes

Importantly, despite a reduction in the risk of macrosomia at birth, the ACHOIS and MFMU follow-up studies did not demonstrate a beneficial impact on childhood obesity and glucose tolerance at 5 to 10 years of age in the offspring of women who received treatment for maternal hyperglycemia (364,365). Other prospective cohort studies similarly suggest that the offspring of women with treated GDM still have a greater risk of obesity, type 2 diabetes, the metabolic syndrome, and cardiovascular disease from early childhood and adolescence (258,366-380). For example, a 2017 Danish National Birth Cohort study (n = 561) reported increased adiposity, an adverse cardiometabolic profile, and earlier onset puberty among adolescent females of women with GDM (381). A prospective offspring cohort study of women with GDM who achieved good antenatal glycemic control demonstrated that offspring adiposity (adipose tissue quantity measured using magnetic resonance imaging) was similar in the GDM and normal glucose tolerance groups within 2 weeks postpartum but was 16.0% greater (95% CI 6.0-27.1; P = 0.002) by 2 months of age (382). The mechanism for this greater adiposity and rapid weight gain in early infancy is uncertain given both groups were predominantly breastfed. Consistent with the ACHOIS and MFMU follow-up studies (364,365), these data suggest that the current approach to glycemic control in GDM may not mitigate its impact on longer term infant health. Further, this pathway may be potentially mediated by excess infant adiposity, which correlates with childhood adiposity (383). Table 8 presents practical tips for managing women with GDM.

Precision Medicine in GDM: Physiological Heterogeneity, Subtype Classification, Risk Prediction, and Biomarker Utility

Precision medicine seeks to improve diagnostics, prognostics, prediction, and therapeutics in diabetes, including GDM, by evaluating and translating various biological axes including metabolomics, genomics, lipidomics, proteomics, technology, clinical risk factors and biomarkers, and mathematical and computer modeling into clinical practice (384). The Precision Medicine in Diabetes Initiative was launched in 2018 by the ADA, in partnership with the European Association for the Study of Diabetes, with their first consensus report published in 2020 (384).

In GDM, precision medicine represents the increasing understanding of heterogeneity within its genotype and phenotype (170,385-388) to identify and translate subclassification of GDM into more personalized clinical care (388). For example, physiologic subtypes of GDM based on the underlying mechanisms leading to maternal hyperglycemia have been recently characterized (386). Among 809 women from the Genetics of Glucose Regulation in Gestation and Growth pregnancy cohort, heterogeneity in the contribution of insulin resistance and deficiency to GDM were characterized based on validated indices of insulin sensitivity and secretory response measured during the 75-g OGTT performed between 24 and 30 weeks’ gestation (388). Compared to women with normal glucose tolerance, women with insulin resistant GDM (51% of GDM) had higher BMI and fasting glucose, hypertriglyceridemia, and hyperinsulinemia, larger infants, and almost double the risk of GDM-associated pregnancy complications. In contrast, women with predominantly insulin secretion defects had comparable BMI, fasting glucose, infant birthweight, and risk of adverse outcomes to those with normal glucose tolerance (388).

Other studies have also suggested that greater insulin resistance in GDM carries a higher risk of perinatal complications (389). A recent multicenter prospective study of 1813 women evaluating subtypes of GDM based on insulin resistance (389) found that women with GDM and high insulin resistance [n = 189 (82.9%)] had a higher BMI, systolic blood pressure, fasting glucose, and lipid levels in early pregnancy compared to women with normal glucose tolerance or those diagnoses with insulin-sensitive GDM. Insulin-sensitive women with GDM [n = 39 (17.1%)] had a significantly lower BMI than women with normal glucose tolerance but similar blood pressure, early pregnancy fasting glucose and lipid levels, and pregnancy outcomes. Despite no differences in insulin treatment and early postpartum glucose intolerance among the GDM subtypes, women with GDM and high insulin resistance had a greater than 2-fold risk of preterm birth and an almost 5-fold increased risk of neonatal hypoglycemia compared with women with normal glucose tolerance. This suggests the high insulin resistance GDM subtype has a greater risk of pregnancy complications potentially arising from the resultant fetal hyperinsulinemia (389).

The contemporary precision medicine approach to GDM also includes the increasing exploration of early pregnancy risk prediction and risk management models (390). The traditional binary clinical risk factor approach to identifying women at high risk in early pregnancy is limited by poor sensitivity and specificity, with studies showing that clinical risk factor-based screening fails to identify 10% to over 30% of women with GDM (391-396). The Pregnancy Outcome for Women with Pre-gestational Diabetes Along the Irish Atlantic Seaboard study found that the prevalence of women with GDM who had no risk factors was low, ranging from 2.7% to 5.4% (397). However, despite the absence of risk factors, these women with GDM had more pregnancy complications than those with normal glucose tolerance (397). Other studies have also reported that women without risk factors diagnosed with GDM have comparable pregnancy outcomes to women with GDM identified as high risk (393). Thus, clinical risk factors alone are not predictive of GDM risk for all women. Although some improvement in the predictive accuracy for GDM is seen in clinical risk scoring approaches (158,398), greater improvement via multivariate risk prediction and mathematical or computer models combining clinical risk factors and biomarkers have been reported in the GDM research setting (154-156,399-403).

Biomarkers are defined as a biological observation that substitutes and ideally predicts the clinically relevant endpoint (ie, GDM) (404). Biomarker discovery and application in the early detection of GDM has become a major research area. However, few biomarkers are specific enough for clinical application (405). Most novel biomarkers with potential utility for the prediction of GDM are involved in pathophysiological pathways related to insulin resistance, dyslipidemia, and type 2 diabetes (402,406) but are frequently mediated by maternal obesity (240,407). Early pregnancy risk prediction models for GDM combining clinical risk factors and biomarkers have included various measures of maternal glucose, lipids, adipokines, inflammatory markers, and pragmatic aneuploidy and preeclampsia screening markers, with model performance (area under the curve) up to 0.91 (153,154,399,402,403,408-416). Limitations to the clinical application of novel biomarkers and model performance include heterogeneity in the testing approach to GDM and cohort characteristics, potential overestimation of model performance due to overfitting of the data to the index study population, the lack of external clinical validation studies, and limited regulatory guidance for validating biomarker assays (405).

The COVID-19 Pandemic and GDM

The COVID-19 pandemic has led to dynamic changes in the testing approach and model of care for women with GDM to minimize the risk of virus transmission and because of decreased clinical capacity. Several temporary pragmatic diagnostic strategies have been suggested as an alternative to the OGTT, including measurement of fasting plasma glucose, random plasma glucose, and HbA1c (417-419). A secondary analysis of 5974 women from the HAPO study (420), reported that the UK, Canadian, and Australian COVID-19–modified diagnostic approaches reduced the frequency of GDM by 81%, 82%, and 25%, respectively. Short-term pregnancy complications in the subgroup of women now with undiagnosed GDM (“missed GDM”) were comparable to women diagnosed with GDM based on the Canadian-modified diagnostic criteria, slightly lower for the UK-modified criteria, but significantly lower for the Australasian Diabetes in Pregnancy Association–modified criteria. While all approaches recommend universal testing, the Australian approach adopts a lower fasting glucose threshold of 4.7 mmol/L to identify women who require an OGTT and does not include HbA1c measurement (420). A retrospective UK study of over 18 000 women sought to define evidence-based recommendations for pragmatic GDM testing during the COVID-19 pandemic (421), reporting that ~5% of women would be identified as GDM based on a random glucose threshold ≥ 8.5 mmol/L (153 mg/dL) at 12 weeks’ gestation and fasting glucose ≥ 5.2 to 5.4 mmol/L (94-97 mg/dL) or HbA1c ≥ 5.7% (39 mmol/mol) measured at 28 weeks’ gestation. Each test predicted some, but not all, obstetric and perinatal complications, lacking the sensitivity of the OGTT for the diagnosis of GDM but overall may provide adequate risk stratification where the OGTT is not feasible (421).

Conclusion

GDM is one of the most common complications of pregnancy and is increasing in global prevalence. Diagnosing GDM is important because perinatal complications and stillbirth risk are reduced by treatment. Despite the benefit of identifying and treating GDM, much of the current (short-term) diagnostic and management approach to GDM remains contentious. These differences confound interpretation and application of trial data, preventing a single standard international approach to GDM.

Recent data indicates near normal birthweight and maternity population outcomes in women with GDM based on modern IADPSG criteria in developed countries, demonstrating that even treatment of “milder” maternal hyperglycemia improves pregnancy outcomes. However, most cases of GDM occur in low- and middle-income countries where perinatal risks are far greater and universal 1-step testing may be more practical. There are limited RCT data to guide diagnosis and management in this setting, and further evidence is urgently needed. In developed countries including the United Kingdom, the main issue arguably does not pertain to women diagnosed with GDM but rather high-risk women who remain unscreened (associated with factors such as lower socioeconomic status and higher BMI) who are at highest risk of stillbirth (74).

The background to the various GDM diagnostic criteria is informative in demonstrating that no approach clearly separates risk groups. It is also now evident that a continuum of risk for GDM exists based on both the timing and degree of maternal hyperglycemia. This underscores the difficulty of defining absolute glucose thresholds at a single timepoint in pregnancy for the diagnosis of GDM and is confounded further by variation in glucose measurement due to preanalytical glucose processing and reproducibility issues. Thus, current diagnostic glucose thresholds for GDM must inevitably reflect compromise and consensus.

A precision medicine approach that recognizes GDM subtype and heterogeneity, enhanced by further research into the genetics of GDM and validation of novel biomarkers and new technologies such as continuous glucose monitoring may improve risk stratification, optimize clinical models of care, and facilitate more individualized and consumer-friendly detection and treatment strategies.

The recent HAPO-FUS data confirming the long-term impact of maternal hyperglycemia on maternal and offspring metabolic health (227,262) highlight an important paradigm shift. The approach to GDM should reflect an evidence base that evaluates diagnostic glucose thresholds and measurement within a framework that includes timing of detection and treatment trials with long-term clinical and health economic outcomes. For example, if the ongoing Treatment of Booking Gestational Diabetes Mellitus trial demonstrates a benefit for early GDM detection and treatment, there are implications for the prevailing diagnostic GDM glucose thresholds in later pregnancy. This is because these thresholds were derived from the risk of perinatal complications in a heterogeneous GDM cohort, which included women who would fulfill early GDM criteria.

Other important areas for research include the evaluation of dietary interventions establishing the optimal carbohydrate threshold in GDM, further clarity on the potential long-term impact of intrauterine metformin on the offspring, as well as the efficacy of preconception and early pregnancy preventive strategies targeting risk factors other than glycemia, such as maternal obesity and GWG. Improved obstetric assessment of placental function, especially in late pregnancy, to inform timing of delivery and identify women at highest risk of stillbirth in GDM is also needed.

The complications of GDM may indeed be greater based on the severity of maternal glycemia and associated vascular risk factors. Nevertheless, the traditional focus on diagnostic criteria and short-term antenatal maternal glucose management fails to address the importance of identifying “milder” (IADPSG-defined) GDM as a risk factor for future maternal and offspring diabetes and CVD risk. It should also be apparent that the increasing prevalence of GDM largely reflects the worsening metabolic health burden including prediabetes and obesity in women of childbearing age. The clinical focus for GDM must therefore urgently shift to early postnatal prevention strategies to decrease the progression from GDM to type 2 diabetes and address longer term maternal and offspring cardiometabolic risk post-GDM via a life course management approach.

Contributor Information

Arianne Sweeting, Department of Endocrinology, Royal Prince Alfred Hospital, Sydney, Australia. Faculty of Medicine and Health, University of Sydney, Sydney, Australia.

Jencia Wong, Department of Endocrinology, Royal Prince Alfred Hospital, Sydney, Australia. Faculty of Medicine and Health, University of Sydney, Sydney, Australia.

Helen R Murphy, Diabetes in Pregnancy Team, Cambridge University Hospitals, Cambridge, UK. Norwich Medical School, Bob Champion Research and Education Building, University of East Anglia, Norwich, UK. Division of Women’s Health, Kings College London, London, UK.

Glynis P Ross, Department of Endocrinology, Royal Prince Alfred Hospital, Sydney, Australia. Faculty of Medicine and Health, University of Sydney, Sydney, Australia.

References