Plasma glucose concentration is the product of insulin secretion and action at any given time. Pregnancy introduces many dynamic physiological variables that alter glucose homeostasis compared to the non-pregnant state. Fasting glucose levels fall in the first trimester and are maintained in the second. Although basal plasma glucose trends are downward with further pregnancy progression, glucose peaks after meals are higher and longer (21, 22). The first trimester is characterized by increased insulin secretion and improved insulin sensitivity under elevated maternal estrogen, resulting in concomitant reductions in glucose concentration. An increased insulin effect in early pregnancy ensures optimal maternal glucose uptake and storage to supply increased energy demands later in pregnancy (22). In contrast to the early pregnancy changes in glucose homeostasis, insulin resistance (IR) increases by 30-70% in the second and third trimesters. The resistance to insulin action is mediated by several placental hormones, especially human placental lactogen (HPL), cortisol, estrogen, and progesterone (23, 24). As a result, maternal glucose uptake in muscle and adipose tissue decreases, and glucose is preferentially shunted across the placenta to optimize fetal growth. Maternal energy requirements are met by endogenous glucose production (gluconeogenesis, glycogenolysis) and via the breakdown of adipose tissue (lipolysis and fatty acid oxidation). Therefore, pregnant mothers are especially prone to develop hyperglycemia in the latter part of pregnancy due to an increase in IR to ensure adequate fetal nutrition. Gluconeogenesis contributes to hyperglycemia in GDM, as shown by the work of Catalano and co-workers (25). They documented that hepatic glucose production in pregnancy is suppressed to a lesser extent in women with GDM when compared to controls with normal glucose homeostasis (80% versus 95%). Increased lipolytic activity in adipose tissue, normally inhibited by insulin and stimulated by HPL, generates free fatty acids that contribute to the resistance of skeletal muscle to insulin action and, in addition, decrease the secretory capacity of the pancreatic β-cells (13).

In a euglycemic pregnancy, the physiological increase in IR is accompanied by a compensatory two-to threefold increase in insulin secretion in women with adequate pancreatic β-cell reserve (21). The appearance of GDM coincides with the period of IR in pregnancy. It is proposed that the major contributor (~80%) to GDM is β-cell dysfunction that is temporarily unmasked by the physiological IR of pregnancy (13). Evidence suggests significant heterogeneity exists in the relative contributions of insulin sensitivity and insulin secretory defects in GDM (26–28). GDM is characterized by a mild degree of hyperglycemia (fasting glucose 5.1- 5.6mmol/L, but <7mmol/L ± a 2hr oral 75-gram glucose tolerance value >7.8-8.5mmol/L, but <11.1mmol/L) (29–32). After delivery of the placenta in women with GDM, the temporary acquired, hormone-induced IR subsides. Maternal insulin sensitivity returns to pre-pregnancy levels within 1-5 days after delivery (33). Risk profiles for adverse pregnancy outcomes vary and are influenced by the degree of hyperglycemia and underlying pathophysiology (28, 34, 35).

Powe et al. aimed to characterize physiologic subtypes of GDM by estimating insulin sensitivity and secretion in 809 women at 24-30 weeks’ gestation (36). Compared to women with normal glucose homeostasis, women with GDM and predominant IR (51% of GDM cohort) had higher BMIs and fasting plasma glucose (FPG) values, significantly higher infant birth weights (z score = 0.57 (−0.01 to 1.37) vs birth weight z-score = 0.03 (−0.53 to 0.52) P = 0.001), and a greater risk of GDM-associated adverse outcomes (57.6 vvs28.2%, P = 0.003). These differences were independent of BMI. Women with predominant insulin secretion defects (30% of GDM cohort) had similar BMIs, FPG’s infant birth weights, and risk of adverse outcomes to the women with normal glucose homeostasis (34). Similarly, research by Benhalmina in the Belgian Diabetes in Pregnancy study showed that GDM women with higher IR represent a more unfavourable metabolic profile with a greater risk of immediate adverse pregnancy outcomes (preterm delivery, labour induction, Caesarean section, and neonatal hypoglycemia and ICU admission) (37).Again, the insulin-sensitive GDM women had similar pregnancy outcomes to normoglycemic women. Early (14 ± 4 weeks) postpartum diabetes and prediabetes rates were similar across the GDM subtypes in this study, albeit much lower (0 and 18%, respectively) than the prevalence of early postpartum diabetes in South Africa (37, 38). Miao and co-workers assessed glucose metabolism and physiological factors potentially contributing to hyperglycemia in 321 Chinese women with GDM six years after delivery. As evaluated by homeostasis model assessment (HOMA), the mechanism of hyperglycemia indicated that women with increased IR were more likely to develop prediabetes, while decreased β-cell function contributed to T2DM development (39). Ageing, elevated body mass index (BMI), and ethnicity are proposed factors contributing to IR in women with GDM (40). In a study conducted in our centre, we showed a high prevalence of overweight and obesity in women with hyperglycemia first detected in pregnancy. Increased BMI was, however, not predictive of early post-delivery T2DM (38). Ignell evaluated glucose homeostasis in a culturally diverse population after GDM and found that women of non-European origin were more insulin resistant than European women (41). The development of T2DM after GDM was higher in the non-European group, but the association between BMI and IR was discordant between Asian and Arab women (39). This implies that factors other than BMI may contribute to IR in women of different ethnic groups and socioeconomic status.

The cytokines leptin and adiponectin, secreted by adipose tissue, play an essential role in IR, T2DM, and obesity. Adiponectin enhances insulin secretion and action and therefore exhibits anti-diabetic and anti-inflammatory effects (42). Adiponectin is negatively correlated with IR and BMI. Leptin inhibits hunger and diminishes fat storage in adipocytes (43). A degree of leptin resistance occurs in T2DM and pregnancy, but the degree in GDM exceeds that of the normoglycemic pregnancy (13, 43). Amino-acid transport across the placenta also contributes to fetal macrosomia in GDM, resulting from increased placental leptin production in hyperinsulinemia and hyperglycemia (44–46).

In non-pregnant individuals, increased BMI may exist in the absence of IR, and the association between adiponectin and IR has been described as independent of a raised BMI (47, 48). In pregnancy, the human placenta is an additional source of adiponectin. Compared to normoglycemic pregnancies, lower adiponectin levels have been documented in GDM, which are associated with decreased insulin sensitivity and reduced anti-inflammatory effects (49). Insufficient placental adiponectin production influences maternal metabolism and contributes to fetal macrosomia (50, 51). In a recent South African study, pregnant women with GDM had higher BMIs, were more insulin resistant, and had lower adiponectin levels than non-GDM women (51). Interestingly, when GDM women in this study were stratified by weight, the degrees of hyperglycemia and IR were similar in the obese and non-obese individuals, with a trend towards higher adiponectin levels noted in the non-obese group. A follow-up study of the same population showed no association between the three single nucleotide polymorphisms (SNPs) within the adiponectin gene that regulates adiponectin expression (ADIPOQ rs266729 and rs17300539 and MTHFR- rs1801133). These SNPs have been associated with an increased risk for GDM in other populations (52).

Auto-immunity against the β-cell is traditionally linked to Type 1 diabetes mellitus (T1DM). Given the high background prevalence of obesity, many patients with a phenotype in keeping with T2DM and IR presently demonstrate autoimmunity. Prevalence figures for autoimmune diabetes (AID) in non-insulin-dependent, adult-onset diabetes range between 3.4-16% (53, 54). Autoimmunity may therefore contribute to the heterogeneity observed in HFDP and GDM. Autoimmunity is noteworthy in this high-risk, vulnerable population due to the high probability of becoming insulin dependent early in the disease course and having persistent hyperglycemia after GDM. Nillsson and colleagues indicated that 6% of women had auto-antibodies indicative of T1DM at the GDM diagnosis (55).The clinical phenotypes of the GDM antibody-positive and antibody-negative women in their cohort (n=385) were clinically indistinct at diagnosis (55). A subset of women diagnosed with GDM (± 10%, auto-antibody positive) are therefore at significantly higher risk (2-4-fold) to develop either classic Type 1 diabetes or latent autoimmune diabetes of adulthood (LADA) six months-10 years postpartum (56, 57). The prevalence of autoimmune antibodies against the β-cell, their role in diabetes development, and their association with HFDP have not been studied in South Africa. It is plausible that the IR of pregnancy unmasks auto-antibody mediated β-cell dysfunction in a subset of South-African women diagnosed with GDM.

In summary, a heterogeneity of physiologic processes underlying hyperglycemia exists among women with GDM. Increased IR appears to confer a greater risk of adverse outcomes in populations from developed countries. Most studies aimed at delineating the pathophysiological processes of GDM have been performed in West-African and Caucasian persons, with South African women grossly under-represented. The relative contribution of β-cell dysfunction, albeit on a polygenetic basis or due to autoimmunity and IR in South-African women with GDM during and after pregnancy, has not been adequately studied. A clearer understanding of these pathogenic pathways and improved knowledge regarding their relative contributions to the development of GDM may lead to more precise management strategies, including new preventative and therapeutic approaches.

The International Association for Diabetes in Pregnancy Study Group (IADPSG) revised the diagnostic criteria for hyperglycemia in pregnancy following the landmark Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study (34, 58). Based on the degree of hyperglycemia, a gradient of risk for antenatal fetal- and maternal adversity was documented. It became increasingly apparent that gestational diabetes lacked precision regarding ante- and perinatal risk stratification (58). Recognizing these limitations, the World Health Organization (WHO) revised the terminology in 2013 (30). The phrase hyperglycemia, first detected in pregnancy, replaced gestational diabetes mellitus as the collective term for all glucose abnormalities first recognized in pregnancy (Figure 1). Based on the degree of hyperglycemia, they also recommended that HFDP be sub-classified into either overt diabetes in pregnancy (DIP) or gestational diabetes mellitus (GDM) (30). On the continuum of elevated glucose concentrations in pregnancy, DIP thus represents greater and GDM milder degrees of hyperglycemia associated with significantly different risk profiles, respectively.

HFDP indicates that the abnormality is first detected in pregnancy, irrespective of the degree and duration. Women with HFDP who meet the WHO criteria for diabetes in non-pregnant individuals (fasting plasma glucose (FPG) ≥7mmol/L ± 2-hour plasma glucose (2-hour PG) following a 75-gram oral glucose tolerance test (OGTT) ≥11.1mmol/L ± an HbA1c value of ≥ 6.5%) at diagnosis, are classified as having DIP (29, 30, 59). Women with HFDP who have milder hyperglycemia (FPG ≥ 5.1-6.9 mmol/L ± 2hr PG ≥ 7.8-11.0 mmol/L following the 75-gram OGTT) have GDM.

Women with overt diabetes first detected in pregnancy, and those with pre-existing known Type 1 diabetes mellitus (T1DM) or Type 2 diabetes mellitus (T2DM) are collectively categorized as DIP (30). With increased awareness regarding the subtypes of diabetes in pregnancy and recent findings of impaired insulin secretion in a subset of GDM women, it has been proposed that the most efficient way to discriminate GDM types is to include testing for pancreatic β-cell antibodies, specifically glutamic acid decarboxylase (GAD) antibody in the assessment of abnormal glucose homeostasis in pregnancy (60). Hyperglycemia in pregnancy thus includes a broad spectrum of glucose abnormalities that range from overt diabetes that predates the index pregnancy to GDM that arises during pregnancy, as depicted in Figure 1.

The one-step 75-gram 2-hour OGTT is the validated diagnostic test for diabetes in the non-pregnant population (59, 83). It is also accepted by most societies globally as the diagnostic test of choice for GDM, albeit with assessment at different time points and with different glycemic cut-offs (30, 58, 70). The ACOG accepts the one-step approach proposed by the former organizations but recommends a two-step approach as an alternative. [Step 1 (50-gram 1-hour) and if screen positive, Step 2 (100-gram 3-hour)] (71).

Even the one-step OGTT is a cumbersome test. Thus, practical alternatives are increasingly pursued, especially in developing countries and healthcare facilities with enormous patient burdens and limited resources, exacerbated by the recent COVID-19 pandemic. The OGTT requires the patient to be fasted overnight to obtain an FPG, then drink a standardized load of 75-gram of glucose, followed by additional blood sampling, thereby spending a minimum of 120 minutes at a health facility. In the non-pregnant state, and when screening for GDM based on NICE criteria, blood sampling is required at two-time points, i.e., at 0 and 120 minutes. A diagnosis of GDM is made if one or both blood glucose levels are in keeping with diagnostic criteria (59, 83). IADPSG guidelines require sampling at three-time points, namely at 0, 60, and 120 minutes and propose a diagnosis of GDM if any of the three predefined glycemic thresholds are met or exceeded (29).

Apart from the considerable time the patient must spend at a health facility, the OGTT also requires human resources and laboratory input. The glucose load ingested as part of the OGTT is, in addition, often associated with unpleasant symptoms such as nausea and vomiting. The need for alternative GDM testing has gained significant attention during the COVID-19 pandemic. To reduce infections, health authorities and professional bodies in the United Kingdom (UK), Canada, and Australia have issued a joint statement to limit the need for pregnant women to attend prolonged appointments, including OGTTs (84). A fasting plasma glucose (FPG) per se has been proposed as an alternative to the OGTT for screening and diagnostic purposes. Only half of the GDM cohort had impaired fasting glucose (FPG ≥5.1mmol/L) in the HAPO study. By contrast, in the other 50% of the cohort, the diagnoses of GDM were based solely on impaired glucose tolerance following the glucose load at 60 minutes ± 120 minutes (66). Recent work from our Diabetes in Pregnancy clinical research team at Tygerberg Hospital has shown that 83% of patients with more overt hyperglycemia and 56% with milder hyperglycemia are diagnosed based on fasting glucose (threshold 5.6mmol/L) (38). In a study by Nachtergaele et al., the sensitivity of 57%. (CI 0.52-0.62) when IADPSG FPG (5.1mmol/L) was compared with OGTT, improved marginally (78% (0.74-0.82)) when the FPG threshold was lowered to 4.6mmol/L (85). Dickson and co-workers compared the utility of an FPG of ≥4.5mmol/L to risk factor-based, selective screening combined with the IADPSG FPG threshold of 5.1mmol/L for diagnosing GDM in an urbanized cohort of black South-African women (85) They demonstrated that universal screening with an FPG determination and an OGTT (that was only performed if FPG ≥4.5mmol/L) had greater sensitivity and specificity (98% and 80%, respectively) in identifying GDM. Their findings proposed a universal screening strategy using FPG ≥4.5mmol/as a more efficient, cost-effective, and resource-effective method to diagnose GDM than risk factor-based selective screening in black South African women (86). Globally, it remains uncertain whether an FPG measurement alone is sensitive enough to screen and diagnose GDM.

To limit the unpleasant gastrointestinal symptoms associated with the 75-gram glucose ingestion, Marais and co-workers compared the 75-gram 2-hour OGTT with a standardized, designed breakfast (87). Overall, the 2-hour PG values of the designed breakfast correlated with the 2hr PG of the 75-gram OGTT (r=0.542, P<0.001), but the small study population did not allow for scrutiny at clinically relevant diagnostic glucose thresholds (87).

Glycated hemoglobin is globally accepted as a useful modality for monitoring and evaluating longer-term control in the non-pregnant population with diabetes mellitus. In addition, an HbA1c ≥6.5% (48 mmol/mol) is regarded as one of the WHO diagnostic criteria for overt diabetes mellitus (59, 83). During pregnancy, a diagnosis of DIP is also made in women who present with HFDP if the HbA1c is ≥6.5% (48 mmol/mol) (30, 58, 59). Discrepancies as to what should be regarded as a normal HbA1c and the ideal HbA1c value for the diagnosis of diabetes remain between different populations and ethnicities (38, 89).

HbA1c in the non-pregnant population is not used to define lesser degrees of impaired glucose homeostasis and is thus not included as a diagnostic criterion for metabolic syndrome per se. Similarly, diagnostic HbA1c thresholds for GDM have not been established, and consensus has not been reached regarding the usefulness of HbA1c measurements to predict or confirm the presence of GDM. Although unsuitable for diagnosing GDM at present, HbA1c is widely used to monitor glucose control in pregnant women with GDM, Type 1, and Type 2 diabetes (30, 59, 83).

The HbA1c test has advantages, as it requires one blood sample and no fasting. All red blood cells contribute to the measured HbA1c, with levels depending on the erythrocyte lifespan, which averages 120 days in the general population. HbA1c levels decrease in pregnancy, reflecting lower maternal glucose and increased red cell turnover in pregnancy. O’Kane et al. proposed the normal reference range of HbA1c in normoglycemic pregnant women to be 4.1–5.9% (21–41 mmol/mol) (91).

Renz and colleagues evaluated the accuracy of HbA1c for GDM diagnosis and observed that while HbA1c has a high specificity, the sensitivity remained low at various thresholds (92). In a meta-analysis consisting of eight studies, the diagnostic accuracy of an HbA1c threshold of 5.7% (39 mmol/mol) had an excellent specificity (95%), which limits the usefulness as a screening test for GDM. Similarly, Nachtergaele et al. reported from France that only 26% (70/266) of women with confirmed GDM on OGTT had an HbA1c ≥5.7% (39 mmol/mol) (85). The combination of HbA1c ≥5.7% (39mmol/mol) with FPG ≥5.1mmol/L or FPG of ≥ 4.6mmol/L improves the diagnostic utility and increases sensitivity to 57% and 78%, respectively. HbA1c levels in a local GDM prevalence study noted the mean HbA1c at 24-28-week gestation to be 5.1% (5.07-5.18%) (32mmol/L; 32-33mmol/L) and 5.3% (5.21-5.39%) (34mmol/mol; 33-35mmol/L) in normoglycemic and GDM women respectively (51). During the COVID-19 pandemic, various organizations proposed HbA1c for GDM screening. The UK and Canadian interim GDM guidelines proposed an HbA1c cut-off of 5.7% (39 mmol/mol) with or without a random ± fasting glucose (84, 93). Due to limited data regarding the association between HbA1c and perinatal outcome for HbA1c values below the diagnostic threshold for overt diabetes, it is not presently regarded as a diagnostic tool for GDM (34).

The utility of HbA1c to distinguish between the various diagnostic categories (IADPSG, NICE) within the GDM spectrum remains to be elucidated. While studies from our group quantified the risk of postpartum diabetes based on antenatal HbA1c at GDM diagnosis, no study in South Africa has focused primarily on the diagnostic utility of HbA1c for GDM (94).

The measurement of HbA1c is influenced significantly by factors that influence red blood cell lifespans, such as iron deficiency or treatment with iron supplements, that affect many women during pregnancy (95). Albumin is the most abundant serum protein, and exposure to glucose leads to the non-enzymatic addition of reducing sugars to its amino acids (glycation). Albumin has a serum half-life of 2 to 3 weeks and a higher glycation rate than hemoglobin (96). Glycated albumin (GA) can thus be used as an intermediate biomarker of glycemic control in non-pregnant women and may reflect changes in the glycemic status of patients earlier than HbA1c. GA may be a practical diagnostic test and potentially provide a more accurate measure of glucose control during pregnancy (97). Caveats include the fact that serum plasma albumin decreases in normal pregnancy, reflecting hemodilution, and, being a component of the acute phase response, decreases in systemic inflammation (98, 99). Furthermore, GA measurements are affected by the presence of proteinuria, and BMI negatively influences GA levels (100). These factors must be considered when using GA as a marker of glycemic status in pregnancy. In a study by Hashimoto et al., GA levels remained stable throughout pregnancy, whereas HbA1c levels decreased during late gestation and were inversely correlated with ferritin levels and transferrin saturation (97). An accurate GA threshold for diagnosing GDM remains elusive, and the application of GA as a screening tool for GDM is therefore controversial (101, 102). The use of GA for GDM diagnosis has not been investigated in pregnant South-African women.

Intermediary metabolism is intricately linked to adipose tissue and mediated by adipose-derived factors or adipokines. Leptin and adiponectin have been discussed earlier (see pathophysiology section). Overall, the adipokines’ association with obesity, inflammation, T2DM, and GDM are well elucidated, but their value is limited by lack of specificity (105).

Markers of inflammation associated with GDM are also involved in the IR of euglycemic pregnancies. IL-6 is a promising biomarker that may improve the early prediction of GDM. Tumour necrosis factor-alpha (TNF-alpha) may indicate IR and contribute to β-cell dysfunction. The association between TNF-alpha and other surrogate markers of inflammation such as c-reactive protein is also strengthened obesity in GDM (106). Collectively, inflammatory cytokines are central to the pathogenesis of GDM, but their utility to diagnose GDM is limited by the lack of distinction between association and causality.

More knowledge on placental factors (pregnancy-associated plasma protein-A (PAPP A), placental growth factor, as well as sex-hormone-binding globulin (SHBG) and their respective roles in GDM, is required to add value to clinical practice (106, 107).

Genetic variants and single nucleotide polymorphisms (SNPs) are significantly associated with T2DM in genome-wide association studies (GWAS) and recently in GDM (108–110). Variants in four maturity-onset diabetes of the young (MODY) genes and the PDX1 gene have also been implicated in GDM development (52, 111). In a study on black South-African women, Botha and colleagues assessed 16 single nucleotide polymorphisms (SNPs) in five MODY and the PDX1 gene. They found a significant inverse association between women with the minor allele of rs4581569 in PDX1 and GDM (111). Dias and co-workers investigated genetic polymorphisms associated with adiponectin expression. They found that ADIPOQrs266729 and rs17300539 and MTHFR rs1801133 polymorphisms are not associated with GDM in a population of black South African women. The author concludes that limited access to genetic testing, the role of gene-gene and gene-environment interactions, and the small study samples limit extrapolation of these findings to clinical practice.

Telomeres, the non-coding caps at the ends of chromosomes, are shortened by oxidative stress and inflammation and are associated with the development of diabetes (112). Many studies report telomere shortening in placentas from pregnancies with uncontrolled diabetes and GDM (113–115). Telomere shortening is also linked with future diabetes and cardiovascular disease in mothers and offspring (116–118). A better understanding of telomere homeostasis in GDM may illuminate some of the mechanisms underpinning health and disease’s developmental origins. Still, again, this has yet to be studied in South-African women with GDM (119, 120).

Epigenetic changes are increasingly implicated in the pathophysiology of metabolic diseases, including GDM (110). Intra-uterine exposure to famine and over-nutrition in the form of hyperglycemia results in an epigenetic change. Epigenetic mechanisms regulating and altering gene expression and function include DNA methylation, microRNA (miRNA), and histone modifications (110, 121, 122). There is tremendous plasticity during development, and in response to environmental factors, DNA methylation is the most thoroughly studied epigenetic mechanism (122). Bouchard and co-workers found that lower DNA adiponectin promoter methylation levels are correlated with higher maternal IR and glucose levels after OGTT during the second and third trimesters in GDM (122). Locally, Dias et al. investigated global DNA methylation in 63 (138 controls) women with GDM and reported the association between obesity and serum adiponectin (123). They noted no difference in DNA methylation between women with GDM and normoglycemic women. Their finding might be because global DNA methylation is too crude to detect subtle but significant differences associated with mild hyperglycemia in pregnancy (HIP) in their population, suggesting the need for gene-specific methylation (123). Subsequent gene-specific methylation analysis by the same authors indicated that a total of 1046 cytosine-phosphate-guanine sites (CpG) (associated with 939 genes) were associated with GDM. DNA methylation of the top five CpG loci showed distinct methylation patterns in GDM and correlated with glucose concentrations. One CpG site mapped to the calmodulin-binding transcription activator 1 (CAMTA1) gene, which has been shown to regulate insulin production and secretion, may offer potential as an epigenetic biomarker in the South-African population (124).

The second most studied epigenetic mechanism, Micro-RNA, has been identified as a significant regulator of β-cell mass, β-cell function, and lipid metabolism (125). Circulating microRNAs are proposed as promising diagnostic biomarkers and predictors of complications in T2DM (124–126). In the last decade, the relevance of microRNAs in GDM has been studied extensively. Zhu et al. identified a microRNA signature composed of five microRNAs (miR-16-5p, miR-17-5p, miR-19a-3p, miR-19b-3p, and miR-20a-5p) were upregulated in the plasma of women affected by GDM (126). Tagoma et al. identified 15 upregulated miRNAs involved in 41 metabolic pathways that included insulin signalling in women with GDM (127). In South-African women with GDM, the expression of two of these microRNA’s, miR-20a-5p and miR-222-3p, was decreased (128). In subsequent work, the same authors studied the expression of these microRNAs in HIV-positive pregnant women with and without GDM (GDM n=15; non GDM n=52). They found no difference in expression between the two cohorts (129).

While the genetic and epigenetic landscape may provide diagnostic biomarkers with clinical utility in GDM, both globally and in South-African women, the clinical use, at present, remains limited. Further validation of these novel biomarkers is needed to elucidate their potential role in women with GDM.

In the short term, ante- and adverse perinatal outcomes include congenital abnormalities, miscarriage (both early), preterm birth, macrosomia or growth restriction, stillbirth, hypertensive disorders of pregnancy, birth trauma, cesarean section, neonatal hypoglycemia, and jaundice (135). These fetal and maternal complications generally correspond to the degree of hyperglycemia, the timing and duration of conception, and organogenesis (136–138). The risks associated with GDM and adverse perinatal outcomes were quantified by the landmark Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study (89). This multicentre, multi-ethnic observational study that included 23316 pregnant women found that maternal hyperglycemia, less severe than diabetes, was associated with limited but important adversity. These adverse effects occurred irrespective of maternal obesity and included hypertensive disorders in the mother, cesarean delivery, fetal adiposity, increased neonatal birth weight (above 90^th centile), raised cord-blood C-peptide levels, and neonatal hypoglycemia. The results of that study formed the basis of the 2010 IADPSG recommendations on the diagnosis and classification of hyperglycemia in pregnancy (58). Their revised diagnostic glucose values for GDM represented values at which infant birth weight, cord C-peptide and percent body fat > 90^th percentile reached 1.75 times the estimated odds of these outcomes, based on fully adjusted logistic regression models. There are limited data on neonatal outcomes after GDM in sub-Saharan Africa, but macrosomia appears to be a common complication (139). Macaulay et al. found no difference between neonatal birth measures in GDM‐exposed and unexposed neonates in Soweto, South Africa (77). Mild hyperglycemia, timely and appropriate treatment, as well as earlier delivery (GDM 38 (37–39) weeks vs non-GDM 39 (38–40) weeks (P = 0.001) might explain this. Interestingly, a gender-specific sub-analysis showed that GDM exposure was associated with increased fetal growth measures, especially abdominal circumference in males, evident at 16–18 weeks’ gestation (77).

Data on the perinatal outcome of patients with pregestational diabetes and GDM attending Groote Schuur Hospital, Cape Town, South Africa, indicate preterm delivery rates of 38.7% for women with T2DM and 34.9% for GDM (140). In that retrospective audit, the perinatal mortality rate was 2.5% (25/1 000 births), with pre-existing diabetes (T1D and T2DM) contributing to most deaths (6.4% and 4.2% within the subgroups, respectively). The overall rate of congenital malformations was 2.4% (n=18 cases), 1.5% for GDM (within generally accepted background incidence), and 4.6% and 5.7% for T2DM and T1D, respectively. In a retrospective audit from the neighbouring tertiary institute in the Western Cape Province, Tygerberg Hospital, lower preterm delivery rates (17.4% and 8.4% in mothers with T2DM and GDM) were recorded (141). Most women in the latter study were overweight or obese (GDM 95%; T2DM: 70%), and chronic hypertension was present in 23% and 42% with GDM and T2DM. Seven cases of fetal anomalies occurred in mothers with T2DM (7.6%), and the combined perinatal mortality rate for T2DM and GDM was 5.3% (52.6/1000 births), with only one early neonatal death (141).

During pregnancy, the variation in glucose metabolism impacts the ability to maintain good glycemic control and predisposes women to medical emergencies such as diabetic ketoacidosis (DKA), associated with fetal mortality rates as high as 27% (142, 143). A pregnant woman oscillates between the fed and fasted state, with the latter characterized by “accelerated starvation.” Alternate fuels are provided for the mother during pregnancy, while glucose is reserved for the fetus (144). Increased IR in pregnancy is interrelated with a perceived intracellular state of starvation, with resultant endogenous glucose production and lipolysis that generates free fatty acids (FFAs) (145, 146). FFAs lead to ketosis and increase maternal IR with decreased insulin secretion based on lipotoxicity through reduced β-cell function (13).

The DKA risk in pregnancy is further amplified by any condition exacerbating insulin demand. A particular danger is that DKA recognition is often delayed in women without known pre-existing diabetes. Schneider et al. noted infection (27%) and omission of insulin therapy (18%) to be the most common precipitants of DKA in 2025 pregnant women in the US (143). DKA without significant elevation in blood glucose levels, i.e., euglycemic DKA, is also more common in pregnancy, a diagnosis often missed or delayed if not considered. Euglycemic DKA represents between 0.8% and 1.1% of all DKA cases in pregnancy (143, 145, 146). The lower-than-expected glucose concentration is attributed to physiologic hemodilution, increased placental glucose transporters, glomerular filtration with glycosuria, and a decrease in glycogenolysis and hepatic glucose production.

A pregnancy complicated by hyperglycemia indicates a significantly increased risk for the mother to have persistent glucose abnormalities after delivery or to develop abnormal glucose homeostasis and even T2DM later in life (147). This risk is dependent on the degree of antenatal hyperglycemia and the duration of follow-up (148). However, even amongst women diagnosed with GDM based on IADPSG criteria, the risk of developing T2DM was increased more than three-fold after 11 years of follow-up (OR=3.4; GDM:52%; non-GDM:20%) and after adjusting for maternal variables (149). Our group demonstrated a high prevalence (26%) of T2DM 6-12 weeks after delivery in women with antenatal HFDP using NICE criteria (38). Chivese et al., in a cross-sectional study performed 5-6 years after delivery on 220 women with GDM in Cape Town, South Africa, indicated that 48% of women had T2DM. At the time of the study, the mean age of the women was 37.2 years (SD 6.0), and 47% of participants were unaware of their diabetes status (150).

The impact of GDM on the increased future risk of cardiovascular (CV) disease, irrespective of the development of T2DM, has recently been assessed in a systematic review and meta-analysis by Kramer, Campbell, and Retnakaran (151). GDM conferred a 2.3-fold increased risk of CV events in the first decade postpartum. Thus, even without developing T2DM, GDM identifies a vulnerable population and provides an opportunity for timely CV risk modification. Data on CV risk in South-African women with HFDP is limited to a single study. Nicolaou et al. documented a ten-fold increased risk to develop T2DM 3-6 years postpartum (4.6 and 27.6-fold for GDM and DIP, respectively) in black African women with HFDP (152). They noted a higher mean carotid intima-media thickness, a surrogate marker for atherosclerotic CVD in women with HFDP compared to normoglycemic pregnant women.

More extensive studies with control groups and longitudinal surveillance programs after GDM are urgently needed to define and mitigate CV risk in South-African women.

In 2017, one in seven live births globally were affected by hyperglycemia in pregnancy (HIP) (20). Several epidemiological studies have since noted detrimental effects of exposure to maternal GDM on adipose tissue distribution and documented an increased risk for T2DM in young offspring exposed to HIP (147, 153). The HAPO follow-up study evaluated glucose metabolism in 4832 ethnically diverse women at 24-28 weeks gestation (149, 154). Higher intra-uterine glucose was linearly associated with offspring glucose levels [maternal FPG with child Impaired Fasting glucose (IFG) and maternal 1-h and 2-h PG with child Impaired Glucose Tolerance (IGT)]. Eleven percent of the offspring of mothers with GDM had IGT compared with 5% of mothers without GDM. The GDM exposed adolescents were insulin resistant (adjusted mean difference -76.3%, CI -130.3 to -22.4) and had limited β-cell compensation as evidenced by an oral disposition index (−0.12, CI −0.17 to −0.064) compared with the offspring of mothers without GDM (149, 154).

In the Exploring Perinatal Outcomes among Children (EPOCH) study, GDM exposure was strongly associated with more significant IR during childhood and adolescence, even after adjustment for BMI (155). This supports the notion held by many clinicians involved with the care of reproductive-aged women that HIP is perpetuating the increasing global burden of obesity, metabolic syndrome, and T2DM. Proposed mechanisms by which in-utero exposure to hyperglycemia influences the risk of T2DM in the offspring point towards epigenetic modifications, changes in insulin receptor expression and function, adipokine metabolism, and β cell development (156–158). DNA methylation is an epigenetic marker that has been implicated in the observed long-lasting metabolic effects following intrauterine exposure to hyperglycemia. Shiau and co-workers published the first study investigating the association between prenatal GDM exposure and DNA methylation in 3–10-year-old offspring in the Tianjin GDM Observational study (159). They found that offspring of mothers with GDM exhibit premature epigenetic ageing by accelerated DNA methylation, which is associated with cardiometabolic risk in the GDM youth compared to controls. Locally, Soepnel and co-workers assessed the association between HFDP (GDM and DIP) and child obesity in pre-school-aged children in Johannesburg, South Africa. The prevalence of overweight/obesity was 10.5% and 11.1% in the 3–6-year-old children exposed to GDM and DIP, respectively, while only 3.9% in the offspring of normoglycemic mothers (160). Fat mass indices were significantly higher in offspring exposed to DIP but attenuated when maternal BMI was adjusted. In another South African study performed in Cape Town, Chivese evaluated the anthropometry of the offspring exposed to HFDP at birth and pre-school age. At birth, 29.6% of the neonates were classified as large for gestational age (LGA), and 12.2% had macrosomia. The prevalence of LGA and macrosomia mirrored the degree of antenatal hyperglycemia, with a higher prevalence in neonates exposed to DIP compared to GDM. The combined prevalence of overweight and obesity was 26.5% at pre-school age. Maternal third trimester 2-hour PG was associated with z-scores for birthweight and pre-school age weight, indicating that this population’s antenatal 2-hour PG was a modifiable risk factor (161).

In a longitudinal study conducted in the PIMA Indian population, maternal hyperglycemia was the strongest predictor for T2DM in youth. Maternal hyperglycemia accounted for 40% of T2DM development in offspring evaluated at age 5-19 years (162). Children exposed to GDM in utero are at risk of being insulin resistant with limited β-cell compensation compared with the offspring of mothers without GDM. The exposure is significantly and independently associated with childhood IGT, occurring over and above the genetic predisposition and persisting after accounting for adiposity (162). This implicates a transgenerational and potential intergenerational effect of exposure to in-utero hyperglycemia on offspring’s metabolic function and CV risk. It remains to be elucidated to what extent and mechanism T2DM development in the offspring is influenced by the degree of antenatal glucose exposure and the epigenetic changes arising in the placenta of the mothers with hyperglycemia.