WHAT IS ALREADY KNOWN ON THIS TOPICWHAT THIS STUDY ADDS

This article introduces novel findings derived from a cohort of women with hyperglycemia in pregnancy within a multi-ethnic population.

It proposes that a unique abnormal fasting-only OGTT phenotype is associated with the Indian ethnicity, while both abnormal fasting-only and 2-hour-only OGTT phenotype are associated with the Chinese ethnicity.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICYIntroduction

Hyperglycemia in pregnancy (HIP) is characterized as any degree of glucose intolerance first identified during pregnancy1 and is associated with significant immediate and long-term adverse maternofetal health outcomes.2–6 The current standard for the diagnosis of HIP is the use of a three-point (fasting, 1 hour and 2 hour) oral glucose tolerance test (OGTT) between 24 and 28 weeks of gestation, based on the guidelines recommended by the International Association of Diabetes and Pregnancy Study Groups (IADPSG),7 based on the Hyperglycemia and Adverse Pregnancy Outcomes (HAPO) multinational study.8

Based on the IADPSG criteria, distinct antenatal OGTT phenotypes have been identified based on abnormal glucose levels observed at fasting, postprandial (1 hour or 2 hour only), or multiple timepoints.9–11 The identification of unique OGTT phenotypes was facilitated by the observation of considerable heterogeneity in plasma glucose (PG) readings among ethnic groups reported in the HAPO cohort of heterogeneous, multinational, ethnically diverse pregnant women.8 For instance, abnormal fasting PG as defined by the IADPSG criteria was found in only 24% of women in Bangkok and 26% of women in the Hong Kong site, whereas it exceeded 70% in Barbados, Bellflower and Providence.12 13 Similarly, abnormal 1-hour OGTT values ranged from 9% in Barbados to 64% in Bangkok, while the 2-hour values exceeded the diagnostic threshold in 6% in Bellflower as compared with 29% in Hong Kong.12 13 In line with these population-level studies, Arora et al also found that Indian women with HIP displayed significantly higher levels of fasting PG as compared with their Swedish counterparts.14 These PG differences may potentially point to distinct OGTT phenotypes among ethnic groups.

OGTT phenotypes are of clinical significance as they have been linked to distinct sets of maternofetal outcomes.9–11 15–18 Uncovering these phenotypes may thus shed light on differing underlying mechanism of glucose dysregulation among ethnicities18 and translate to differential responses to various treatment modalities.19

Despite its clinical importance, the applicability of OGTT phenotypes across various ethnicities, especially Southeast Asian ethnicities, remains uncertain. In our local multi-ethnic Asian population, only one study has looked into ethnic-specific OGTT phenotypes in postpartum women.20 In this study, the authors categorized patients with HIP into three postnatal OGTT phenotypes based on the number of timepoints with glucose abnormalities (one, two or three timepoints) and found that Indian women had glucose abnormalities at more timepoints, as compared with their Malay and Chinese counterparts.20 There still remains a paucity of research into the antenatal OGTT phenotypes, specifically in women with HIP.

Our study aims to address this significant gap by examining the association between three different Asian ethnic groups (Chinese, Malay and Indian) and unique OGTT phenotypes defined by the proportions of exclusively abnormal 0-hour, 1-hour, and 2-hour PG timepoints in those diagnosed with HIP in a multi-ethnic Asian population. We hypothesize that each ethnic group will be associated with a unique OGTT phenotype.

Materials and methodsSubjects

This cross-sectional study was conducted using retrospective data from women diagnosed with HIP. Briefly, hospital data were extracted from the SingHealth Electronic Health Intelligence System (eHINTs) for 7495 pregnant women who were seeking antenatal care for their index pregnancy at KK Women’s and Children’s Hospital (KKH), one of the largest public maternity units in Singapore between January and December 2019.

Among 7495 women screened, a total of 4367 subjects met one or more of the exclusion criteria. We excluded women with pre-existing diabetes (n=124), multiple pregnancies (n=139), those diagnosed with hyperglycemia in pregnancy (HIP) before 24 weeks or after 28 weeks (n=3779) and those who were of ethnicities other than Malay, Chinese, or Indian (n=325). We also excluded cases with missing data on HIP diagnosis (n=19) and incomplete ethnicity information (n=82), leaving a sample of 3027 women with complete data (online supplemental table 1). Of the 3027 women who underwent a 75 g, 3-point OGTT between 24 and 28 weeks’ gestation,21 16.8% of them (508 cases) were diagnosed with HIP using the International Association of Diabetes and Pregnancy Study Groups (IADPSG) criteria at 24–28 weeks: fasting glucose ≥5.1 mmol/L, 1-hour glucose ≥10.0 mmol/L, and 2-hour glucose ≥8.5 mmol/L.7 The final analytic sample consisted of 508 women diagnosed with HIP, predominantly Chinese (47.6%), followed by Malays (36.8%) and Indians (15.6%). These included subjects were from homogeneous parental ethnic backgrounds (ie, both parents from the same ethnic group). All women diagnosed with HIP were managed according to standard routine care.22 23 Figure 1 depicts the inclusion and exclusion criteria used to select the study population.

Figure 1

Subjects included for analysis in a study of ethnic differences in HIP OGTT phenotypes. DM, diabetes mellitus; HIP, hyperglycemia in pregnancy; T1DM, type 1 diabetes mellitus; T2DM, type 2 diabetes mellitus.

Data collectionOGTT phenotypes

Unique OGTT phenotypes were defined based on the possible combination of abnormal PG values at fasting (≥5.1 mmol/L), 1 hour (≥10.0 mmol/L), and 2 hours (≥8.5 mmol/L) based on the IADPSG criteria.7 The OGTT phenotypes examined in the study were classified into four mutually exclusive groups based on abnormal PG at only single timepoints or combination timepoints: (1) 0 hour only, (2) 1 hour only, (3) 2 hour only, (4) ≥2 or more timepoints (0 hour+1 hour, 0 hour+2 hours, 1 hour+2 hours, 0 hour+1 hour+2 hours).

Covariates

Maternal age, parity (first, second, or third child onward), and ethnicity (Chinese, Malay or Indian) information were obtained via self-report at the time of recruitment. First trimester body mass index (BMI) was calculated using each woman’s measured height (cm) and their measured weight (kg) on the day of their booking scan, between 0 and 12 weeks’ gestation. BMI was categorized as underweight, normal weight, overweight and obese (<18.5, 18.5–23, 23–27.5, ≥27.5 kg/m2) according to the WHO Asian BMI criteria.24 Gestational weight gain (GWG) was calculated based on the difference between the measured weight at the last obstetric visit and the booking visit, assuming a term birth and a first-trimester booking visit. Excessive GWG thresholds, as defined by Gong et al, were set at 10.30 kg for underweight, 10.14 kg for normal weight, and 6.26 kg for overweight or obese BMI categories.25 Gestational age was calculated based on self-reported last menstrual period at the time of the booking visit. Baseline height and weight for women with a late booking visit (>12 weeks) were excluded from the analysis.

Statistical analyses

Continuous variables are presented as mean values with SD, and categorical variables are reported as frequencies and percentages. Univariate analyses were performed to compare the demographic and clinical characteristics of the three main ethnic groups (Chinese, Malays, and Indians). Continuous and categorical independent variables were analyzed using the Kruskal-Wallis test and the χ2 test, respectively. Post hoc analyses were conducted via Dunn’s pairwise comparisons and χ2 adjusted residuals for continuous and categorical variables, respectively.

Adjusted multivariate analyses were undertaken using a multinomial logistic regression with robust variances to estimate the relative risk ratios (RRRs) between the ethnic-specific relative risk ratios and OGTT phenotypes, accounting for maternal age, parity, and first-trimester BMI as potential confounders. Multiple imputations were used for missing BMI information (n=658), employing sequential chained equations. The missing BMI data were found to be missing completely at random (MCAR) using Little’s MCAR test (Chi-statistic of 0.46, p=0.50). We conducted two sets of sensitivity analyses. We restricted the analyses to a subset in subjects with complete BMI information (n=396). We also assessed the association between ethnicity and OGTT phenotypes which consisted of a (1) fasting (0 hour only) phenotype and a (2) post-load (1-hour-only or 2-hour-only) phenotype using a multinomial logistic regression with robust variances analysis, adjusting for the same confounders as before. A p value <0.05 was considered statistically significant. All point estimates were presented with 95% CIs. Statistical analyses were conducted using STATA software V.17SE (StataCorp LLC, College Station, Texas, USA).

Results

The maternal demographic and clinical characteristics of these subjects are summarized in table 1. The mean gestational age at observation of first-trimester BMI was 7.7 weeks (SD 2.13). The Chinese women were older, more likely to be primiparous, and less likely to be obese compared with Malay and Indian women. The mean serum glucose level across the entire cohort at the fasting, 1-hour and 2-hour timepoints are 4.82, 10.6 and 8.46 mmol/L, respectively. Table 1 further illustrates the mean serum glucose level at each OGTT timepoint for each of the ethnicities.

Table 1

Maternal demographics of HIP subjects by ethnic groups

Ethnic-specific OGTT phenotypes

Figure 2 illustrates ethnic variations in OGTT phenotypes within the study cohort in only the women who developed HIP. Of the 3027 women who underwent the OGTT, 508 (16.8%) were diagnosed with HIP. Ethnic-specific rates demonstrated the highest prevalence in Indians (20.5%), followed by Chinese (18.3%) and Malays (14.2%) (online supplemental table 1).

Figure 2

Venn diagrams depicting ethnic-specific oral glucose tolerance phenotypes.

There were considerable variations in the proportions of those with abnormal PG values at specific single timepoints among the three ethnic group. At the fasting-only timepoint, the highest proportion was the Indians (16.5%), followed by Chinese (8.3%), and Malays (5.9%). At the 1-hour only timepoint, Malays were most likely to surpass the threshold (34.2%), followed by Chinese (33.5%) and Indians (24.1%). At the 2-hour-only timepoint, Chinese were most likely to surpass the threshold (23.6%), followed by Indians (10.1%) and Malays (9.6%) (figure 2).

Overall, 48.0% of subjects exhibited abnormal PG values at combination timepoints. Of these, Malay individuals represented the largest proportion (38.5%), with Indian subjects closely following (34.4%) and Chinese participants comprised 16.0% (figure 2).

The association between ethnicity and unique OGTT phenotypes

Table 2 presents the multivariable-adjusted model of the association between specific ethnic groups and unique OGTT phenotypes at the 0-hour, 1-hour or 2-hour timepoints. Compared with Malays, the Chinese and Indians had a significantly higher risk of having an abnormal fasting OGTT phenotype with a RRR of 2.33 (95% CI 1.01 to 5.38, p<0.05) and RRR of 3.05 (95% CI 1.23 to 7.56, p<0.05), respectively, with Indians being at the highest risk after adjusting for confounders (table 2).

Table 2

Associations between ethnic groups and unique OGTT phenotypes at 0-hour, 1-hour and 2-hour timepoints

Only the Chinese women were associated with an increased risk of displaying an OGTT phenotype with abnormal glucose tolerance threshold at 2-hour only with an RRR of 2.88 (95% CI 1.51 to 5.47, p<0.01) as compared with Malay women. There were no significant associations between the ethnic groups and the risk of having an OGTT phenotype characterized by an abnormal 1-hour plasma glucose.

Discussion

To our knowledge, this is the first study to identify unique ethnic-specific OGTT phenotypes in women diagnosed with HIP in Southeast Asia. Specifically, we found that Indian women had a higher risk of having an abnormal fasting OGTT phenotype, while Chinese women were at higher risk of displaying an OGTT phenotype with abnormal fasting and 2-hour-only glucose abnormalities.

Several studies have demonstrated that a noteworthy portion of individuals with HIP in the Indian population manifest abnormal fasting glucose levels.26 Gopalakrishnan et al found that 70.5% of the HIP cases in their study exclusively exhibited abnormal fasting glucose levels.26 Similarly, Moradi et al noted that 48% of the HIP cases exclusively presented with elevated fasting blood glucose.27 While our study only found that 16.5% of Indian women had exclusively abnormal fasting glucose, this proportion was still higher than Chinese (8.3%) and Malay (5.9%) women. In line with our findings, two other studies have shown that non-pregnant Asian Indians exhibit a higher rate of fasting glucose abnormalities as compared with Caucasians.28 29 To date, only one other study in pregnant women has reported findings that align with our own. A study by Arora et al concluded that Punjabi Indian women with HIP exhibited significantly higher levels of fasting glucose compared with their Swedish counterparts.14 This pattern of fasting dysglycemia in Indians is further corroborated by the observation that stricter fasting glucose thresholds resulted in more distinctive cases of HIP (n=90) diagnosed in a South Asian Sri Lankan population as compared with when stricter 2-hour glucose thresholds were used (n=15).30

In contrast to the fasting dysglycemia observed in Indians,28 29 postprandial dysglycemia has been reported in individuals of Chinese ethnicity.31 32 This distinction aligns with our observations, where Chinese women demonstrated a distinct OGTT phenotype characterized by abnormalities in post-load glucose levels. Wang et al found a higher postprandial serum glucose increment in Chinese non-pregnant individuals as compared with those of white European ancestry.31 Similarly, Simper et al noted that, despite having similar fasting blood glucose levels, Chinese non-pregnant healthy individuals showed a higher postprandial glucose incremental area under the curve than their white European counterparts.32 Notably, our study extends these observations to pregnant Chinese women diagnosed with HIP, marking the first report of this pattern in this demographic.

Unique OGTT phenotypes among the ethnic groups may allude to ethnic-specific differences in insulin resistance. Specifically, two forms of insulin resistance have been identified in the literature: (1) hepatic insulin resistance, marked by the liver’s reduced ability to suppress gluconeogenesis in response to insulin, measured by the Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) index, and (2) peripheral insulin resistance, primarily affecting muscle cells, evaluated through the insulin clamp technique, which assesses insulin’s efficacy in promoting glucose uptake.26 33 In our study, Indian women were associated with a unique fasting OGTT phenotype, which may allude to an underlying hepatic insulin resistance.18 In contrast, we observed that Chinese women had a post-load OGTT phenotype, which may be due to peripheral insulin resistance.18 A review noted that fasting glucose abnormalities are associated with hepatic insulin resistance, while postprandial glucose abnormalities are characteristic of peripheral insulin resistance.33 Insulin resistance patterns have been observed in pre-diabetic obese adolescents, showcasing three distinct phenotypes: impaired fasting glucose (IFG), impaired glucose tolerance (IGT), and the combined IFG/IGT phenotype.34 Individuals exhibiting IGT and IFG/IGT demonstrated peripheral insulin resistance, whereas IFG was associated with heightened hepatic insulin resistance during the second phase of insulin secretion.34 These differences in glucose metabolism has important implications on treatment modality of HIP.

Unique OGTT phenotypes uncovered in this study might indicate the need for ethnic-specific treatment modalities of HIP.15–17 As previous studies have shown, individuals with fasting glucose abnormalities are associated with increased insulin therapy, while those with postprandial glucose abnormalities had reduced insulin usage and responded to diet and exercise therapy.17 19 Other than differences in treatment modalities, distinct OGTT phenotypes have potential clinical implications on both immediate and long-term pregnancy outcomes. Current evidence suggests distinct sets of immediate maternofetal outcomes associated with OGTT phenotypes.35–38 Women who surpassed the glucose threshold only at the fasting timepoint displayed higher pre-conceptional BMI,37 39 required more frequent insulin therapy,37–39 and were at higher risk of having a primary cesarean section.37 These women were also at higher risk of impaired postpartum dysglycemia.18 The children of these women were born with a significantly higher mean birth weight and were at higher risk of being large for gestational age.36–38 In contrast, women who exceeded only at post-load timepoints were significantly more likely to have gestational hypertension,36 38 preterm delivery,36 emergent C-section,37 and were at an increased risk of having children who were small for gestational age or were delivered with low fetal weight (<30th percentile).37 In terms of long-term adverse outcomes, previous research has also used OGTT phenotypes for risk stratification. This study demonstrated a dose-response relationship, indicating that women who failed more timepoints during the OGTT exhibited a progressively increased risk of developing postpartum dysglycemia.20 Taken together, the evidence highlights the practical potential use of ethnic-specific OGTT phenotypes based on both type (fasting vs post-load) and quantity (number of timepoints) for diagnosis, risk stratification and tailored management of HIP.

The study’s major strength lies in its utilization of the dataset from KKH, a major tertiary obstetrics center in Singapore, which encompasses three major Asian ethnic subgroups that make up a large proportion of the world population. Our dataset is extensive and comprehensive, providing a rich source for analysis that includes patient demographics and a complete set of 3-point OGTT results. Nevertheless, as this is the first study of its kind, our study numbers may still be insufficiently powered to detect significant differences among ethnicities. Our findings are also limited to the specific population examined at KKH and are not generalizable to other populations. Future studies with larger, more diverse populations will be needed to validate our findings and explore their broader applicability. The retrospective design limits our ability to draw causal inferences, and there is a critical need for future prospective cohort studies to provide more robust evidence regarding the relationship between ethnicity and OGTT phenotypes in women with HIP. As with any retrospective data analysis, causality cannot be established without corroborating evidence. Furthermore, the influence of residual confounding cannot be completely ruled out resulting from data that was not collected (ie, patient’s country of origin, immigration status, diet and physical activity status) or not well collected (ie, family history or obstetric history). Lastly, since our hospital medical records did not adequately capture the HIP treatments patients were receiving, resulting in a significant amount of missing data, we were unable to determine how the OGTT results among the different ethnic groups affect HIP treatment.

In conclusion, our research alludes to the presence of ethnic-specific OGTT phenotypes with Indians being at higher risk of abnormalities in fasting glucose, whereas Chinese were at higher risk for abnormalities in both the fasting at 2-hour glucose values. There is a need for future studies to disentangle the possible underlying mechanisms associated with these unique OGTT phenotypes.

Data availability statement

Data are available upon reasonable request.

Ethics statementsPatient consent for publication

Not applicable.

Ethics approval

Ethical approval was obtained from the SingHealth Centralised Institutional Review Board for an exempt review (CIRB Reference No. 2019/2510) as this is a retrospective study of deidentified data.

Acknowledgments

We would like to acknowledge the KKH data management team for their support in extracting the necessary data for this project.