Associations of fasting plasma glucose and glycosylated hemoglobin levels at less than 24 weeks of gestation with hypertensive disorders of pregnancy: the BOSHI study

Abstract

This study aimed to evaluate the associations of fasting plasma glucose (FPG) and glycosylated hemoglobin (HbA1c) levels at <24 weeks of gestation with hypertensive disorders of pregnancy (HDP) and compare the strengths of the associations of HDP with FPG and HbA1c levels. Totally, 1,178 participants were included in this prospective cohort study. HDP, FPG, HbA1c, and potential confounding factors were included in multiple logistic regression models. The number of HDP cases was 136 (11.5%). When FPG and HbA1c were included in the model separately, quartile 4 (Q4) of FPG (87–125 mg/dL) and HbA1c (5.2–6.3% [33–45 mmol/mol]) levels had higher odds of HDP than quartile 1. The odds ratios (ORs) were 1.334 (95% confidence interval [CI]: 1.002–1.775) for Q4 of FPG and 1.405 (95% CI: 1.051–1.878) for Q4 of HbA1c. When the participants were divided into two categories based on the cut-off value with the maximum Youden Index of FPG or HbA1c, the ORs for high FPG (≥84 mg/dL) or high HbA1c (≥5.2% [33 mmol/mol]) were 1.223 (95% CI: 1.000–1.496) and 1.392 (95% CI: 1.122–1.728), respectively. When both FPG and HbA1c were included in the model simultaneously, the statistical significance of Q4 of FPG disappeared, whereas that of HbA1c remained. In two-category models, the same results were obtained. High FPG and HbA1c levels at <24 weeks of gestation were risk factors for HDP in pregnant Japanese women. In addition, high HbA1c levels were more strongly associated with HDP than high FPG levels.

1.  Introduction

Hypertensive disorders of pregnancy (HDP) are adverse perinatal outcomes and risk factors for eclampsia, maternal and neonatal death, premature delivery, and fetal growth restriction [1]. Therefore, diagnosis of HDP at prenatal checkups is important for early treatment intervention and intensive care in tertiary hospitals.

Hyperglycemia during pregnancy is a common complication and remains a risk factor for HDP [2]. The Hyperglycemia and Adverse Pregnancy Outcome (HAPO) study demonstrated the associations between high fasting plasma glucose (FPG) and glycosylated hemoglobin (HbA1c) levels during 24–32 weeks of gestation with preeclampsia (PE), a subtype of HDP [3, 4]. Hyperglycemia during pregnancy is generally diagnosed using a 75-g oral glucose tolerance test (OGTT) during 24–28 weeks of gestation based on the results of the HAPO study. However, the diagnostic criteria differs depending on the country or organization because of the lack of evidence to diagnose hyperglycemia during early pregnancy [5]. Consequently, the association between the glucose metabolism status at <24 weeks of gestation and perinatal outcomes needs to be investigated [6-8]. Both FPG and HbA1c levels are used to detect hyperglycemia at <24 weeks of gestation [9, 10]. Even when the HbA1c level is <6.5% (<47 mmol/mol), the higher the HbA1c level, the higher the risk of HDP and other adverse pregnancy outcomes [11-13]. Nevertheless, the association between FPG level at <24 weeks of gestation and HDP remains unclear and has been inadequately investigated in pregnant Japanese women [14-16]. In addition, whether FPG or HbA1c is more strongly associated with HDP remains unclear. Identifying risk factors for HDP remains crucial, as the administration of low-dose aspirin (LDA) during early pregnancy in high-risk women could reduce occurrence of preterm PE. While both type 1 and type 2 diabetes mellitus are recognized risk factors for PE, and LDA use is recommended by the American College of Obstetricians and Gynecologists (ACOG) [17] and the Fetal Medicine Foundation (FMF) predictive model [18] for PE, hyperglycemia during early pregnancy is not accounted for in the predictive model for PE. The absence of consensus on diagnosing and intervening in hyperglycemia during early pregnancy and the unclear impact of intervention may contribute to its exclusion from the predictive model for PE.

Thus, this study aimed to assess the associations between FPG and HbA1c levels at <24 weeks of gestation with HDP among pregnant Japanese women. Additionally, it aimed to compare the relative strengths of these associations with HDP.

2.  Materials and Methods

2.1  Study design

This study was a part of the “Babies and Their Parents Longitudinal Observation in Suzuki Memorial Hospital on Intrauterine period” (BOSHI) study, a prospective cohort study conducted at Suzuki Memorial Hospital, an obstetrical and gynecological hospital, in the Sendai city area, Miyagi Prefecture, Japan. The Suzuki Memorial Hospital is a regional hospital that provides care primarily for low-risk pregnancies. All study protocols were approved by the Institutional Review Board of Tohoku University School of Medicine (No. 2019-7) and the Hospital Review Board of Suzuki Memorial Hospital. Written informed consent was obtained from all participants. Details of the BOSHI study are described elsewhere [19, 20].

2.2  Measurement of FPG and HbA1c

Maternal venous blood was collected after overnight fasting at <24 weeks of gestation, after informed consent was obtained from the participants. Both FPG and HbA1c levels, the exposures in this study, were measured by a subcontractor (SRL, Inc., Tokyo, Japan). The FPG levels were assayed using the hexokinase ultraviolet method (CicaLiquid GLU [EM] Reagent1; KANTO KAGAKU, Tokyo, Japan). The intra- and inter-assay coefficients of variation (CV) were 1.48% and 0%, respectively. The detection threshold was 2 mg/dL. The HbA1c levels were assayed using an enzymatic method (MetaboLead HbA1c; Minaris Medical, Tokyo, Japan). The intra- and inter-assay CV were 2.91% and 0%, respectively, and the detection threshold was 3.3% (National Glycohemoglobin Standardization Program [NGSP]). HbA1c levels were defined according to the NGSP, but they were recorded as the Japan Diabetes Society (JDS) value until April 1, 2012. Therefore, we converted the HbA1c level (JDS), measured until April 1, 2012, into HbA1c level (NGSP), using the following formula: HbA1c (NGSP) in % = HbA1c (JDS) in % * 1.02 + 0.25 [21]. We also described the HbA1c levels based on the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) using the following formula: IFCC HbA1c in mmol/mol = 10.93 * NGSP HbA1c in % – 23.52 [22].

2.3  Definition of HDP

Clinic blood pressure (CBP) was measured at each prenatal checkup. The method for measuring CBP has been previously described [20]. HDP, an outcome of this study, was diagnosed based on the CBP levels. The diagnosis of HDP was based on the Japan Society for the Study of Hypertension in Pregnancy guidelines [23]. HDP was defined as hypertension (CBP ≥140/90 mmHg) between 20 weeks of gestation and 12 weeks postpartum. HDP was subclassified into gestational hypertension, PE, superimposed preeclampsia, and eclampsia. Gestational hypertension was defined as persistent hypertension without proteinuria. PE was defined as new-onset hypertension and proteinuria (≥300 mg per day) between 20 weeks of gestation and 12 weeks postpartum. As this study focused on new-onset hypertension during pregnancy, participants with chronic hypertension and superimposed preeclampsia were excluded.

2.4  Data collection

Data on maternal age, height, pre-pregnancy body weight (BW), parity, smoking status during pregnancy, alcohol intake, HDP in prior pregnancy, family history of hypertension, and family history of diabetes mellitus (DM) were collected using questionnaires at the first visit. Data on maternal BW on the day of blood sampling; conception method, including assisted reproductive technology (ART); gestational age at delivery; infant birth weight; and infant sex were collected from medical charts. Maternal hemoglobin (Hb) value at <24 weeks of gestation was also collected.

Pre-pregnancy body mass index (BMI) was calculated as pre-pregnancy BW in kilograms divided by the square of maternal height in meters. Gestational weight gain was calculated by subtracting the pre-pregnancy BW from maternal BW on the day of blood sampling. The seasons at the time of blood sampling were defined as follows: spring, March–May; summer, June–August; autumn, September–November; and winter, December–February.

2.5  Statistical analysis

All statistical analyses were conducted using SAS software Ver 9.4 (SAS Institute Inc., Cary, NC, USA). Continuous variables are presented as mean ± standard deviation (SD) or median (range), as appropriate. Categorical variables are expressed as numbers and percentages.

A multiple logistic regression model was applied to investigate the associations of FPG and HbA1c levels with HDP at <24 weeks of gestation. Model 1 was defined as the crude model. Model 2 was adjusted for maternal age, pre-pregnancy BMI, gestational weight gain, HDP in prior pregnancy, family history of hypertension, family history of DM, smoking status during pregnancy, alcohol intake, parity (primipara or not), conception by ART, season at the time of blood collection, and Hb levels at <24 weeks of gestation [19, 24, 25]. With reference to the previous study that both FPG and HbA1c levels also exhibit seasonality, the season at the time of blood collection, rather than the season at the time of conception or the season at delivery was adjusted in our study [25]. Model 3 was adjusted for mean arterial pressure (MAP) measured for the first time after conception (i.e., initial MAP) in addition to the variables in Model 2, because the BP level during early gestation might be a confounding factor [26]. A meta-analysis indicated that MAP is a better predictor of PE than systolic blood pressure (SBP) or diastolic blood pressure (DBP); therefore, MAP was included in Model 3 [27]. MAP was calculated using the following formula: MAP = DBP + (SBP – DBP)/3. The MAP at the time of collection of maternal venous blood was not included in the model because it was an intervening variable. No strong multicollinearity among the explanatory variables was confirmed based on the variance inflation factor using a general linear model with HDP as a continuous variable.

First, FPG and HbA1c levels were included separately in a multiple logistic regression model. The study participants were divided into quartiles of FPG at <24 weeks of gestation, and the lowest quartile (i.e., quartile 1) was set as the reference category. We also divided the participants into two categories based on the cutoff value with the maximum Youden index [28]. Similarly, the study participants were divided into quartiles and two categories based on their HbA1c levels. In addition, the linear trend association of the quartiles of FPG or HbA1c levels with HDP was tested. FPG and HbA1c levels were also included in the model separately as continuous variables per 1-SD increase. Second, both FPG and HbA1c levels were included in multiple logistic regression models simultaneously to compare the strengths of the associations of FPG and HbA1c levels with HDP at <24 weeks of gestation. Next, both FPG and HbA1c levels were simultaneously included in the models in three ways: quartile (i.e., four categories), two categories based on cutoff values that had the maximum Youden index, and continuous variables per 1-SD increase. A likelihood ratio test was performed to compare the goodness of fit between the two models (FPG vs. FPG + HbA1c). Additionally, the goodness of fit between the two models (HbA1c vs. HbA1c + FPG) was compared using the likelihood ratio test.

As several explanatory variables had missing data, we used multiple imputation when a multiple logistic regression model was used. The missing explanatory variables were both continuous and categorical, and the missing pattern was non-monotonic. Therefore, we applied multiple imputation using a chained equation (MICE). After 10 datasets were created using MICE, the same analysis was performed for each dataset. The results have been combined and reported in the manuscript. The likelihood ratio test after multiple imputation was performed using the SAS COMBCHI macro [29]. Statistical significance was defined as a two-sided p-value <0.05.

3.  Results

3.1  Maternal and neonatal characteristics

The flowchart of this study is shown in Fig. 1. A total of 1,450 pregnant women at <24 weeks of gestation consented to participate in this study between October 16, 2006, and October 7, 2011. Thirteen women who withdrew consent, 11 with twin pregnancies, 31 women whose FPG and HbA1c levels were not measured at <24 weeks of gestation, and four women who were diagnosed with overt diabetes in pregnancy were excluded. Twenty women with chronic hypertension, eight women who developed superimposed preeclampsia, 17 women who experienced stillbirth, and two women diagnosed with primary aldosteronism were excluded. Furthermore, 91 women with missing data on the development of HDP were excluded, along with 75 women who were transferred to other hospitals for GDM management, relocation, or other reasons. Finally, 1,178 Participants who gave birth at the Suzuki Memorial Hospital were analyzed.

Fig. 1  Flow chart of this study

HbA1c, glycosylated hemoglobin

The maternal and neonatal characteristics are shown in Table 1. The maternal age was 30.9 ± 4.9 (mean ± SD) years. The median gestational age at the time of blood collection was 13.0 (range 9.4–23.4) weeks. The FPG value was 83.1 ± 6.4 (mean ± SD) mg/dL. The HbA1c level (NGSP/IFCC) was 5.0 ± 0.3% (31 ± 3 mmol/mol) (mean ± SD). The initial MAP was 81 ± 10 (mean ± SD) mmHg. The number of cases of HDP was 136 (11.5%). The range of FPG levels was as follows: 63–79 mg/dL, quartile 1; 80–82 mg/dL, quartile 2; 83–86 mg/dL, quartile 3; and 87–125 mg/dL, quartile 4. The range of HbA1c levels in each quartile was as follows: 3.2–4.8% (11–28 mmol/mol), quartile 1; 4.9% (30 mmol/mol), quartile 2; 5.0–5.1% (31–32 mmol/mol), quartile 3; and 5.2–6.3% (33–45 mmol/mol), quartile 4.

Table 1 Maternal and neonatal characteristics of study participants

Variables	All N = 1,178
Maternal characteristics
Age, mean ± SD	30.9 ± 4.9
<35 years, n (%)	887 (75.3)
≥35 years, n (%)	291 (24.7)
Pre-pregnancy BMI, kg/m2, mean ± SD	21.7 ± 3.4
<18.5 kg/m2, n (%)	144 (12.2)
18.5–24.9 kg/m2, n (%)	878 (74.5)
≥25.0 kg/m2, n (%)	156 (13.2)
Gestational weight gain, kg, mean ± SD	0.58 ± 2.4
Primipara, n (%)	674 (57.2)
ART, n (%)	28 (2.4)
Family history of diabetes mellitus, n (%)
Yes	155 (13.1)
No	895 (76.0)
Missing data	128 (10.9)
Family history of hypertension, n (%)
Yes	401 (34.0)
No	651 (55.3)
Missing data	126 (10.7)
HDP in prior pregnancy, n (%)	37 (3.1)
Smoking status, n (%)
No smoking before conception	967 (82.1)
Until conception was recognized	163 (13.8)
Smoking during pregnancy	48 (4.1)
Alcohol intake, n (%)
No alcohol intake before conception	593 (50.3)
Until conception was recognized	573 (48.7)
Alcohol intake during pregnancy	12 (1.0)
Gestational age at the time of blood collection, median (range)	13.0 (9.4–23.4)
FPG at <24 weeks of gestation, mean ± SD	83.1 ± 6.4
HbA1c (NGSP, IFCC) at <24 weeks of gestation, %, mean ± SD	5.0 ± 0.3, 31 ± 3
Initial clinic blood pressure level measured for the first time after conception, mmHg
SBP, mean ± SD	110 ± 12
DBP, mean ± SD	66 ± 10
MAP, mean ± SD	81 ± 10
Clinic blood pressure level at the time of blood collection, mmHg
SBP, mean ± SD	109 ± 11
DBP, mean ± SD	67 ± 9
MAP, mean ± SD	81 ± 9
Season at the time of blood collection, n (%)
Spring	308 (26.2)
Summer	311 (26.4)
Autumn	230 (19.5)
Winter	329 (27.9)
HDP, n (%)	136 (11.5)
Gestational hypertension, n (%)	112 (9.5)
Preeclampsia, n (%)	24 (2.0)
Delivery week, weeks	39.7 ± 1.3
Preterm delivery, n (%)	30 (2.6)
Neonatal characteristics
Sex (male), n (%)	589 (50.5)
Birth weight, g, mean ± SD	3,078 ± 396
Low birth weight infants (<2,500 g), n (%)	72 (6.1)
Macrosomia (≥4,000 g), n (%)	11 (0.9)

Data are shown as mean ± SD, median (range), or number (percentages).

Abbreviations: ART, assisted reproductive technology; BMI, body-mass index; DBP, diastolic blood pressure; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin; HDP, hypertensive disorders of pregnancy; IFCC, International Federation of Clinical Chemistry and Laboratory Medicine; MAP, Mean blood pressure; NGSP, National Glycohemoglobin Standardization Program; SBP, systolic blood pressure; SD, standard deviation

Supplementary Tables 1 and 2 show the maternal and neonatal characteristics of the study participants according to the quartiles of FPG and HbA1c levels at <24 weeks of gestation, respectively. The numbers of GDM cases based on the results of 75-g OGTT were not be determined in our study. As shown in the Supplementary Table 3, the number and percentage of participants with both FPG levels of ≥92 mg/dL at <24 weeks of gestation and at ≥24 weeks of gestation was 11 (1.1%)

3.2  Association of FPG level at <24 weeks of gestation with HDP

The association between FPG levels at <24 weeks of gestation and HDP is shown in Fig. 2. Compared with the participants in quartile 1, those in quartile 4 had significantly higher odds of HDP in Models 1, 2, and 3. The adjusted odds ratios (ORs) in quartile 4 were 1.472 (95% confidence interval [CI]: 1.118–1.938) in Model 2 and 1.334 (95% confidence interval [CI]: 1.002–1.775) in Model 3. Quartiles 2 and 3 were not associated with HDP in Models 1, 2, and 3. p-values for the trend were as follows: 0.001, Model 1; 0.013, Model 2; and 0.088, Model 3.

Fig. 2  Associations of FPG and HbA1c levels at <24 weeks of gestation with HDP

^a the cut off value based on Youden index

^b 1-SD values were 6.4 mg/dL for FPG and 0.26% for HbA1c

^c Model 2: Adjusted for maternal age, pre-pregnancy BMI, gestational weight gain, hemoglobin, HDP in prior pregnancy, family history of hypertension, family history of diabetes mellitus, smoking status during pregnancy, alcohol intake, parity (primipara or not), ART, and season at the time of blood collection

^d Model 3 plus initial MAP

Abbreviations: ART, assisted reproductive technology; BMI, body-mass index; CI, confidence interval; FPG, fasting plasma glucose; HDP, hypertensive disorders of pregnancy; IFCC, International Federation of Clinical Chemistry; MAP, mean arterial pressure; NGSP, National Glycohemoglobin Standardization Program; OR, odds ratio

*Statistically significant (p-value <0.05)

The FPG level with the maximum Youden index was 84 mg/dL. Participants with an FPG level of ≥84 mg/dL had significantly higher odds than those with FPG levels <84 mg/dL. The adjusted ORs for participants with FPG levels ≥84 mg/dL were 1.274 (95% CI: 1.051–1.545) in Model 2 and 1.223 (95% CI: 1.000–1.496) in Model 3. When FPG was included in the model as a continuous variable per 1-SD increase, the adjusted ORs for HDP were 1.190 (95% CI: 1.000–1.415) in Model 2 and 1.134 (95% CI: 0.945–1.360) in Model 3.

3.3  Association of HbA1c level at <24 weeks of gestation with HDP

The association between HbA1c levels at <24 weeks of gestation and HDP is shown in Fig. 2. Participants in quartile 4 (HbA1c level of 5.2–6.3% [33–45 mmol/mol]) had significantly higher odds of HDP, compared to those in quartile 1 in Models 1, 2, and 3. The adjusted ORs for participants in quartile 4 were 1.476 (95% CI: 1.120–1.945) in Model 2 and 1.405 (95% CI: 1.051–1.878) in Model 3. p-values for the trend were as follows: <0.001, Model 1; 0.010, Model 2; and 0.027, Model 3. Quartiles 2 and 3 of HbA1c were not associated with HDP in Models 1, 2, and 3.

The HbA1c level that had the maximum Youden index was 5.2% (33 mmol/mol). Participants with HbA1c levels ≥5.2% had significantly higher odds than those with levels <5.2%. The adjusted ORs for participants with HbA1c levels ≥5.2% were 1.439 (95% CI: 1.171–1.769) in Model 2 and 1.393 (95% CI: 1.122–1.728) in Model 3. When HbA1c, as a continuous variable per 1-SD increase, was included in the model, the adjusted ORs of HDP were 1.191 (95% CI: 0.975–1.455) in Model 2 and 1.133 (95% CI: 0.922–1.394) in Model 3.

3.4  Comparison of FPG and HbA1c levels at <24 weeks of gestation in terms of HDP

A comparison of the associations of FPG and HbA1c levels at <24 weeks of gestation in terms of HDP in Model 3 is shown in Fig. 3. When both FPG and HbA1c, as quartiles, were included in the models simultaneously, the adjusted ORs of quartile 4 were 1.260 (95% CI: 0.940–1.689) for FPG and 1.361 (95% CI: 1.014–1.828) for HbA1c in Model 3. The likelihood ratio test revealed an improved goodness of fit when HbA1c was included in the model in addition to FPG (p-values = 0.007) (Supplementary Table 4). When both FPG and HbA1c, divided into two categories based on the Youden index, were included in the models simultaneously, the adjusted ORs were 1.192 (95% CI: 0.972–1.461) for FPG and 1.367 (95% CI: 1.101–1.695) for HbA1c. The likelihood ratio test revealed an improved goodness of fit when HbA1c was included in the model in addition to FPG (p-values = 0.005) (Supplementary Table 5).

Fig. 3  Comparison of associations of FPG and HbA1c levels at <24 weeks of gestation in terms of HDP

^a The cut-off value based on Youden index

^b 1-SD values were 6.4 mg/dL for FPG and 0.26% for HbA1c

^c Both FPG and HbA1c levels at <24 weeks of gestation were included in the model simultaneously. Adjusted for initial MAP, maternal age, pre-pregnancy BMI, gestational weight gain, hemoglobin, HDP in prior pregnancy, family history of hypertension, family history of diabetes mellitus, smoking status during pregnancy, alcohol intake, parity (primipara or not), ART, and season at the time of blood collection

Abbreviations: ART, assisted reproductive technology; BMI, body-mass index; CI, confidence interval; FPG, fasting plasma glucose; HDP, hypertensive disorders of pregnancy; MAP, mean arterial pressure; OR, odds ratio; SD, standard deviation

*Statistically significant (p-value <0.05)

Neither FPG nor HbA1c was associated with HDP when both FPG and HbA1c, as per a 1-SD increase, were included in the models simultaneously. The goodness of fit did not improve in each model when FPG and HbA1c, as per the 1-SD increase, were included in the likelihood ratio test (Supplementary Table 6).

4.  Discussion

To the best of our knowledge, this is the first study to show an association between FPG levels at <24 weeks of gestation and HDP in pregnant Japanese women without overt diabetes in pregnancy. High FPG and HbA1c levels at <24 weeks of gestation were risk factors for HDP in pregnant Japanese women. In addition, a higher HbA1c level was more strongly associated with HDP than a higher FPG level at <24 weeks of gestation.

The results of the association between FPG levels and HDP in this study are in line with those of previous studies. An FPG level of ≥110 mg/dL at <20 weeks of gestation is a risk factor for HDP [14]. An FPG level of ≥80 mg/dL during the first trimester is also a risk factor for PE [15]. However, Sesmilo et al. [16] reported that FPG levels during the first trimester were not associated with HDP. The inconsistency between the results of our study and those of Sesmilo’s study may be due to differences in maternal baseline characteristics and explanatory variables. In their study, the participants had a high mean maternal age of 34.2 years, and a high proportion of nulliparity (88.2%). In addition, their explanatory variables (maternal age, BMI, weeks of gestation when glycemia was observed, previous pregnancies, weight gain during pregnancy, and tobacco use) differed from those in our study. The participants in their study would be considered a higher-risk participants for HDP compared to those in our study and therefore, the strength of the association between FPG and HDP may have differed.

The result of the association between HbA1c levels and HDP in our study is consistent with those of previous studies. A total of 16,122 pregnant women in New Zealand with an HbA1c of 5.9–6.4% (40–46 mmol/mol) during early pregnancy had a significantly higher risk of PE compared to those with HbA1c levels <5.9% [11]. Pregnant women with an HbA1c level of 5.9–6.4% in the first trimester had a significantly higher risk of PE compared to those with an HbA1c level <5.9% [15]. In an ongoing birth cohort study in Japan, the adjusted OR per 1% increase in HbA1c level at <24 weeks of gestation was 1.77 (95% CI: 1.48–2.12) for HDP [13].

Our results of the association of FPG and HbA1c levels at <24 weeks of gestation with HDP were consistent with those of another study [15]. FPG levels were not a better predictor of pregnancy complications (macrosomia, preterm delivery, cesarean section, infant birth weight, and PE) than HbA1c levels [15]. A non-graded association was found between FPG levels >80 mg/dL and PE, but this was not a clinically useful cut-off point. In contrast, HbA1c levels beyond the 5.9% (41 mmol/mol) cut-off level were associated with a three-fold increased risk of PE.

The decisive reason for the HbA1c level during early pregnancy being a better predictor of the development of HDP than the FPG level is unclear. Theoretically, FPG measures blood glucose levels at a given time point, whereas HbA1c reflects long-term glucose control. In the general population, FPG appears to underestimate the burden of undiagnosed diabetes [30]. HbA1c levels may reflect large fluctuations in blood glucose levels during pregnancy [31].

The point estimates of the adjusted ORs for HDP in participants with FPG levels ≥87 or ≥84 mg/dL were high, and therefore, we do not discourage the measurement of FPG at the prenatal checkup. We suggest that medical institutions select between the measurement of FPG or HbA1c at <24 weeks of gestation to identify pregnant women at high risk of HDP by considering the necessity for fasting state, medical costs, and medical insurance coverage.

Although the etiology of HDP has not yet been clearly established, a two-stage disorder theory has been suggested as a mechanism for the development of HDP [32]. Reduced placental perfusion (stage I) due to factors such as abnormal trophoblast invasion and vascular remodeling of the spiral artery is thought to be the cause of the clinical manifestations of PE (stage II). Cawyer et al. [33] reported that hyperglycemia upregulates soluble endoglin, soluble fms-like tyrosine kinase-1 (sFlt-1), and interleukin-6 levels and downregulates placental growth factor (PlGF) and vascular endothelial growth factor levels, which induces anti-angiogenic milieu and inflammation [34]. Additionally, hyperglycemia inhibited cytotrophoblast invasion and induced an angiogenic imbalance, which may explain the association between hyperglycemia during early pregnancy and HDP.

It remains uncertain whether screening for Hyperglycemia during early pregnancy and implementing interventions for blood sugar regulation lead to improved pregnancy outcomes [35, 36]. However, identifying risk factors for HDP is crucial as initiating LDA during early pregnancy in high-risk women may lower the incidence of preterm PE, a subtype of HDP [17, 37]. Further investigation into glucose metabolism status (including FPG and HbA1c) during early pregnancy and its association with PE could help identify pregnant women at heightened risk of developing PE.

This study had some limitations. First, the sample size was not sufficient to evaluate the association between FPG levels and HDP subtypes, such as gestational hypertension and PE. In addition, four pregnant women with early-onset HDP who needed termination were excluded from this study because they were transferred to other hospitals that could provide care for preterm labor. Twelve women (8.8%) in 136 HDP cases had the clinical blood pressure exceeding diagnostic criteria for HDP at less than 34 weeks of gestation in this study. An HbA1c of 5.9–6.4% (40–46 mmol/mol) was a risk factor for adverse pregnancy outcomes, including HDP [10, 12, 14]. However, there were only two pregnant women (0.17%) with an HbA1c ≥5.9% (40 mmol/mol) at <24 weeks of gestation in this study. Therefore, we could not evaluate the association of an HbA1c level of ≥5.9% (40 mmol/mol) with HDP. Information on the implementation of the 75-g OGTT was not collected in this study. Therefore, we could not examine the association between the results of the 75-g OGTT at <24 weeks of gestation and HDP. Our study used data from more than 10 years ago. In Japan, the diagnostic criteria for GDM were revised in 2010 based on the results of HAPO study, and the impact of this change on the results of this study is unclear. Results in our study could not draw a definitive conclusion regarding whether or not to intervene for GDM during early pregnancy because information on intervention for GDM during early pregnancy was not collected. Both observational and randomized controlled studies on the diagnosis and intervention for GDM during early pregnancy, considering FPG and HbA1c levels, are needed in Japan. As shown in Supplementary Table 7, predictive ability for HDP using FPG or HbA1c levels at <24 weeks of gestation was limited in our study. Therefore, it is necessary to combine FPG or HbA1c levels at <24 weeks of gestation with other predictor variables for the improvement of the predictive ability for HDP. We could not evaluate the association between hyperglycemia and angiogenic imbalance because neither sFlt-1 nor PlGF levels were measured in this study.

In conclusion, high FPG and HbA1c levels at <24 weeks of gestation are risk factors for HDP. Moreover, HDP shows a stronger association with high HbA1c levels compared to high FPG levels. Further research on the association between hyperglycemia and HDP is needed to clarify the appropriate diagnoses and interventions for hyperglycemia during early pregnancy to prevent HDP.

 Acknowledgements

We would like to acknowledge the study participants and contributions of the members of the BOSHI study described in the Supplementary Material. We would also like to thank Editage (https://www.editage.com/) for proofreading the manuscript.

 Authors’ Contribution Statement

SI performed statistical analyses and wrote the manuscript. NI contributed to the conception of the study, interpreted the data, and revised the manuscript. HH contributed to the interpretation of data and revised the figure design. TO, MI, and KH contributed to data collection and interpretation and revised the manuscript. M Sato, TM, M Saito, TS, SK, NY, and YI contributed to data interpretation and revised the manuscript. HM supervised the BOSHI study, contributed to data collection, interpreted the data, and revised the manuscript.

 Disclosure

 Conflict of Interest

Omron Healthcare Co., Ltd. was not involved in the study design, statistical analysis, or the writing of the manuscript. Takashi Sugiyama is a member of Endocrine Journal’s Editorial Board.

 Funding

This study was supported by Grants for Scientific Research (18590587, 18390192, 21390201, 24689061, 25253059, 26860412, 16H05243, 17K15857, 18K15837, 19H03905, and 21K10438) from the Ministry of Education, Culture, Sports, Science and Technology of Japan; a Grant-in-Aid (19DA1001) for Health Research on Children, Youth and Families, and (H21-Junkankitou [Seishuu]-Ippan-004) from the Ministry of Health, Labour and Welfare, Health and Labour Sciences Research Grants, Japan; Grant (JP19gk0110039, JP24gn0110088) from AMED; a Grant-in-Aid for Japan Society for the Promotion of Science (JSPS) fellows (19.7152), and JSPS KAKENHI (JP19K18659). Additionally, academic contributions were received from Pfizer Japan Inc.; Bayer Academic Support; Takeda Research Support; Astellas Research Support, J&J Medical Research Grant and scholarship donations from Chugai Pharmaceutical Co., Ltd., Daiichi Sankyo Co., Ltd., and Otsuka Pharmaceutical Co., Ltd.

Supplementary Table 1 Maternal and neonatal characteristics of study participants according to quartiles of FPG level at <24 weeks of gestation.

Maternal and neonatal characteristics of study subjects according to quartiles of FPG level at <24 weeks of gestation (Supplementary Table 1) The higher the FPG level, the higher the proportion of HDP. In addition, the higher the FPG level, the higher the maternal age, prepregnancy BMI, and clinic blood pressure at the time of blood collection.
Characteristics	FPG level at <24 weeks of gestation
	Quartile 1 (63–79 mg/dL) N = 334	Quartile 2 (80–82 mg/dL) N = 262	Quartile 3 (83–86 mg/dL) N = 313	Quartile 4 (87–125mg/dL) N = 269
Maternal characteristics
Age, mean ± SD	30.6 ± 5.0	30.6 ± 4.6	31.2 ± 5.2	31.2 ± 4.7
<35 years, n (%)	251 (75.2)	208 (79.4)	223 (71.3)	205 (76.2)
≥35 years, n (%)	83 (24.9)	54 (20.6)	90 (28.8)	64 (23.8)
Pre-pregnancy BMI, kg/m2, mean ± SD	20.8 ± 2.5	21.4 ± 2.9	21.8 ± 3.7	23.0 ± 3.9
<18.5 kg/m2, n (%)	53 (15.9)	30 (11.5)	37 (11.8)	24 (8.9)
18.5–24.9 kg/m2, n (%)	259 (77.5)	200 (76.3)	233 (74.4)	186 (69.1)
≥25.0 kg/m2, n (%)	22 (6.6)	32 (12.2)	43 (13.7)	59 (21.9)
Gestational weight gain, kg, mean ± SD	0.94 ± 2.5	0.42 ± 2.2	0.54 ± 2.5	0.33 ± 2.5
Primipara, n (%)	186 (55.7)	152 (58.0)	182 (58.2)	154 (57.3)
ART, n (%)	7 (2.1)	6 (2.3)	12 (3.8)	3 (1.1)
Family history of diabetes mellitus, n (%)
Yes	45 (13.5)	32 (12.2)	40 (12.8)	38 (14.1)
No	256 (76.6)	204 (77.9)	234 (75.7)	201 (74.7)
Missing data	33 (9.9)	26 (9.9)	39 (12.5)	30 (11.2)
Family history of hypertension, n (%)
Yes	105 (31.4)	84 (32.0)	113 (36.1)	99 (36.8)
No	196 (58.7)	151 (57.6)	162 (51.8)	142 (52.8)
Missing data	33 (9.9)	27 (10.4)	38 (12.1)	28 (10.4)
HDP in prior pregnancy, n (%)	9 (2.7)	9 (3.5)	8 (2.6)	11 (4.1)
Smoking status, n (%)
No smoking before conception	271 (81.1)	223 (85.1)	250 (79.9)	223 (82.9)
Until conception was recognized	46 (13.8)	30 (11.4)	50 (15.9)	37 (13.7)
Smoking during pregnancy	17 (5.1)	9 (3.5)	13 (4.2)	9 (3.4)
Alcohol intake, n (%)
No alcohol intake before conception	188 (56.3)	123 (47.0)	149 (47.6)	133 (49.4)
Until conception was recognized	146 (43.7)	135 (51.5)	158 (50.5)	134 (49.8)
Alcohol intake during pregnancy	0 (0)	4 (1.5)	6 (1.9)	2 (0.8)
Gestational age at the time of blood collection, median (range)	13.4 (10.0–22.0)	13.3 (10.0–21.9)	13.0 (9.4–22.3)	13.2 (10.0–23.4)
FPG at <24 weeks of gestation, mean ± SD	76.6 ± 2.6	81.1 ± 0.8	84.4 ± 1.1	91.7 ± 6.2
HbA1c (NGSP, IFCC) at <24 weeks of gestation, %, mean ± SD	5.0 ± 0.2, 31 ± 2	5.0 ± 0.3, 31 ± 3	5.1 ± 0.2, 32 ± 2	5.1 ± 0.3, 32 ± 3
Initial clinic blood pressure level measured for the first time after conception, mmHg
SBP, mean ± SD	107 ± 11	110 ± 13	110 ± 12	113 ± 13
DBP, mean ± SD	63 ± 9	65 ± 10	67 ± 9	68 ± 10
MAP, mean ± SD	78 ± 9	80 ± 10	81 ± 10	83 ± 10
Clinic blood pressure level at the time of blood collection, mmHg
SBP, mean ± SD	107 ± 10	108 ± 11	109 ± 11	112 ± 12
DBP, mean ± SD	65 ± 8	66 ± 8	68 ± 9	70 ± 9
MAP, mean ± SD	79 ± 8	80 ± 8	82 ± 9	84 ± 9
Season at the time of blood collection, n (%)
Spring	92 (27.5)	75 (28.6)	77 (24.6)	64 (23.8)
Summer	75 (22.5)	73 (27.9)	94 (30.0)	69 (25.7)
Autumn	75 (22.5)	47 (17.9)	55 (17.6)	53 (19.7)
Winter	92 (27.5)	67 (25.6)	87 (27.8)	83 (30.9)
HDP, n (%)	25 (7.5)	30 (11.5)	36 (11.5)	45 (16.7)
Gestational hypertension, n (%)	23 (6.9)	28 (10.7)	28 (9.0)	33 (12.3)
Preeclampsia, n (%)	2 (0.6)	2 (0.8)	8 (2.6)	12 (4.5)
Delivery week, weeks, mean ± SD	39.7 ± 1.2	39.7 ± 1.2	39.6 ± 1.4	39.8 ± 1.4
Preterm delivery, n (%)	8 (2.4)	6 (2.3)	9 (2.9)	7 (2.6)
Neonatal characteristics
Sex (male), n (%)	180 (53.9)	125 (48.1)	154 (50.2)	130 (48.9)
Birth weight, g, mean ± SD	3,057 ± 367	3,047 ± 396	3,092 ± 391	3,118 ± 435
Low birth weight (<2,500 g), n (%)	18 (5.4)	21 (8.1)	18 (5.8)	15 (5.6)
Macrosomia (≥4,000 g), n (%)	1 (0.3)	5 (1.9)	2 (0.6)	3 (1.1)

ART, assisted reproductive technology; BMI, body mass index; DBP, diastolic blood pressure; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin; HDP, hypertensive disorders of pregnancy; IFCC, International Federation of Clinical Chemistry and Laboratory Medicine; MAP, mean arterial pressure; NGSP, National Glycohemoglobin Standardization Program; SBP, systolic blood pressure; SD, standard deviation

Supplementary Table 2 Maternal and neonatal characteristics of study participants according to quartiles of HbA1c level at <24 weeks of gestation

Maternal and neonatal characteristics of study subjects according to quartiles of HbA1c level at <24 weeks of gestation (Supplementary Table 2) The higher the HbA1c level, the higher the proportion of HDP. The higher the HbA1c level, the higher were the maternal age and pre-pregnancy BMI. The proportions of preterm deliveries and low-birth-weight infants were higher in participants in quartile 4 than in those in quartiles 1–3.
Characteristics	HbA1c level at <24 weeks of gestation
	Quartile 1 3.2–4.8% (11–28 mmol/mol) N = 310	Quartile 2 4.9% (30 mmol/mol) N = 183	Quartile 3 5.0–5.1% (31–32 mmol/mol) N = 396	Quartile 4 5.2–6.3% (33–45 mmol/mol) N = 289
Maternal characteristics
Age, mean ± SD	29.9 ± 5.1	30.4 ± 4.9	30.8 ± 4.7	32.6 ± 4.7
<35 years, n (%)	254 (82.0)	148 (80.9)	302 (76.3)	183 (63.3)
≥35 years, n (%)	56 (18.0)	35 (19.1)	94 (23.7)	106 (36.7)
Prepregnancy BMI, kg/m2, mean ± SD	21.3 ± 2.8	21.4 ± 2.9	21.8 ± 3.4	22.3 ± 3.8
<18.5 kg/m2, n (%)	31 (10.0)	32 (17.4)	49 (12.4)	32 (11.1)
18.5–24.9 kg/m2, n (%)	250 (80.7)	128 (70.0)	293 (74.0)	207 (71.6)
≥25.0 kg/m2, n (%)	29 (9.3)	23 (12.6)	54 (13.6)	50 (17.3)
Gestational weight gain, kg, mean ± SD	0.88 ± 2.7	0.70 ± 2.4	0.51 ± 2.5	0.26 ± 2.0
Primipara, n (%)	205 (66.1)	109 (59.6)	232 (58.6)	128 (44.3)
ART, n (%)	9 (2.9)	6 (3.3)	6 (1.5)	7 (2.4)
Family history of diabetes mellitus, n (%)
Yes	36 (11.6)	22 (12.0)	48 (12.1)	49 (17.0)
No	250 (80.7)	147 (80.3)	301 (76.0)	197 (68.1)
Missing data	24 (7.7)	14 (7.7)	47 (11.9)	43 (14.9)
Family history of hypertension, n (%)
Yes	97 (31.3)	65 (35.5)	138 (34.8)	101 (35.1)
No	189 (61.0)	105 (57.4)	213 (53.8)	143 (49.7)
Missing data	24 (7.7)	13 (7.1)	38 (11.4)	44 (15.2)
HDP in prior pregnancy, n (%)	4 (1.3)	9 (4.9)	8 (2.6)	13 (4.5)
Smoking status, n (%)
No smoking before conception	250 (80.6)	143 (78.1)	328 (82.8)	246 (85.1)
Until conception was recognized	45 (14.5)	32 (17.5)	53 (13.3)	33 (11.4)
Smoking during pregnancy	15 (4.9)	8 (4.4)	15 (3.9)	10 (3.5)
Alcohol intake, n (%)
No alcohol intake before conception	180 (58.1)	121 (66.1)	236 (59.6)	183 (63.4)
Until conception was recognized	125 (40.5)	61 (33.3)	154 (38.9)	105 (36.5)
Alcohol intake during pregnancy	4 (1.4)	1 (0.6)	6 (1.5)	1 (0.3)
Gestational age at the time of blood collection, median (range)	13.3 (10.0–22.4)	12.9 (10.1–22.9)	12.9 (10.0–23.4)	12.7 (9.4–19.4)
FPG at <24 weeks of gestation, mean ± SD	81.7 ± 6.4	82.4 ± 5.8	83.3 ± 6.4	84.8 ± 6.5
HbA1c (NGSP, IFCC) at <24 weeks of gestation, %, mean ± SD median(IQR)	4.7 ± 0.2, 27 ± 3	4.9 ± 0, 30 ± 0	5.0 ± 0.1, 31 ± 1	5.4 ± 0.2, 35 ± 2
Initial clinic blood pressure level measured for the first time after conception, mmHg
SBP, mean ± SD	108 ± 12	111 ± 12	110 ± 12	111 ± 12
DBP, mean ± SD	64 ± 10	66 ± 10	66 ± 9	67 ± 10
MAP, mean ± SD	78 ± 10	81 ± 10	81 ± 9	82 ± 10
Clinic blood pressure level at the time of blood collection, mmHg
SBP, mean ± SD	108 ± 12	110 ± 11	109 ± 11	110 ± 12
DBP, mean ± SD	64 ± 10	68 ± 9	67 ± 8	68 ± 9
MAP, mean ± SD	80 ± 8	82 ± 9	81 ± 9	82 ± 10
Season at the time of blood collection, n (%)
Spring	77 (24.8)	43 (23.5)	101 (25.5)	87 (30.1)
Summer	72 (23.2)	51 (27.9)	117 (29.6)	71 (24.6)
Autumn	69 (22.3)	39 (21.3)	73 (18.4)	49 (17.0)
Winter	92 (29.7)	50 (27.3)	105 (26.5)	82 (28.3)
HDP, n (%)	26 (8.4)	18 (9.8)	38 (9.6)	54 (18.7)
Gestational hypertension, n (%)	20 (6.5)	14 (7.7)	32 (8.1)	46 (15.9)
Preeclampsia, n (%)	6 (1.9)	4 (2.2)	6 (1.5)	8 (2.8)
Delivery week, weeks, mean ± SD	39.7 ± 1.4	39.5 ± 1.3	39.8 ± 1.2	39.5 ± 1.3
Preterm delivery, n (%)	8 (2.6)	4 (2.2)	6 (1.5)	12 (4.2)
Neonatal characteristics
Sex (male), n (%)	164 (53.4)	78 (41.5)	209 (52.8)	138 (47.8)
Birth weight, g, mean ± SD	3,036 ± 403	3,047 ± 396	3,092 ± 391	3,097 ± 388
Low birth weight (<2,500 g), n (%)	21 (6.8)	11 (6.0)	17 (4.3)	23 (8.0)
Macrosomia (≥4,000 g), n (%)	3 (1.0)	1 (0.5)	5 (1.3)	2 (0.7)

ART, assisted reproductive technology; BMI, body mass index; DBP, diastolic blood pressure; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin; HDP, hypertensive disorders of pregnancy; IFCC, International Federation of Clinical Chemistry and Laboratory Medicine; MAP, mean arterial pressure; NGSP, National Glycohemoglobin Standardization Program; SBP, systolic blood pressure; SD, standard deviation

Supplementary Table 3 The numbers and percentages of participants according to both FPG levels at <24 weeks of gestation and at ≥24 weeks of gestation.

		FPG level at <24 weeks of gestation
		<92 mg/dL (N = 960)	≥92 mg/dL (N = 84)
FPG level at ≥24 weeks of gestation	<92 mg/dL (N = 991)	918 (87.9%)	73 (7.0%)
	≥92 mg/dL (N = 53)	42 (4.0%)	11 (1.1%)

FPG, fasting plasma glucose

Supplementary Table 4 The results of the likelihood ratio test when FPG and HbA1c levels were divided into quartiles

Models	Δ (vs. Model N) (Degrees of freedom)
		Δ (vs. each model)	Degrees of freedom	p-value
Model A FPG (Quartiles)	4.421 (3)	4.421a	3	0.036
Model B HbA1c (Quartiles)	8.836 (3)	8.836a	3	0.003
Model C FPG (Quartiles) + HbA1c (Quartiles)	11.70 (6)	7.288b	3	0.007
		2.872c	3	0.091

Δ = –2Log Likelihood (model k) – [–2Log Likelihood (model k + α)]

^a Relative to Model N.

^b Relative to Model A.

^c Relative to Model B.

Model N was a multiple logistic regression model including maternal age, pre-pregnancy BMI, gestational weight gain, hemoglobin, HDP in prior pregnancy, family history of hypertension, family history of diabetes mellitus, smoking status during pregnancy, alcohol intake, parity (primipara or not), ART, season at the time of blood collection, and initial MAP.

Model A was a multiple logistic regression model with FPG (quartiles) in addition to Model N.

Model B was a multiple logistic regression model with HbA1c (quartiles) in addition to Model N.

Model C was a multiple logistic regression model including both HbA1c (quartiles) and FPG (quartiles) levels in addition to Model N.

ART, assisted reproductive technology; BMI, body mass index; DBP, diastolic blood pressure; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin; HDP, hypertensive disorders of pregnancy; MAP, mean arterial pressure

Supplementary Table 5 The results of the likelihood ratio test when FPG and HbA1c levels were divided into two categories based on the Youden index.

Models	Δ (vs. Model N) (Degrees of freedom)
		Δ (vs. each model)	Degrees of freedom	p-value
Model A FPG (≥84 mg/dL vs. <84 mg/dL)	3.860 (1)	3.861a	1	0.049
Model B HbA1c (≥5.2% vs. <5.2%)	8.832 (1)	8.832a	1	0.003
Model C FPG (≥84 mg/dL vs. <84 mg/dL) + HbA1c (≥5.2% vs. <5.2%)	11.69 (2)	7.830b	1	0.005
		2.858c	1	0.091

Δ = –2Log Likelihood (model k) – [–2Log Likelihood (model k + α)]

^a Relative to Model N.

^b Relative to Model A.

^c Relative to Model B.

Model N was a multiple logistic regression model including maternal age, pre-pregnancy BMI, gestational weight gain, hemoglobin, HDP in prior pregnancy, family history of hypertension, family history of diabetes mellitus, smoking status during pregnancy, alcohol intake, parity (primipara or not), ART, season at the time of blood collection, and initial MAP.

Model A was a multiple logistic regression model with FPG (two categories based on the Youden index) in addition to Model N.

Model B was a multiple logistic regression model with HbA1c (two categories based on the Youden index) in addition to Model N.

Model C was a multiple logistic regression model including both HbA1c (Youden index) and FPG (Youden index) levels in addition to Model N.

ART, assisted reproductive technology; BMI, body mass index; DBP, diastolic blood pressure; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin; HDP, hypertensive disorders of pregnancy; MAP, mean arterial pressure

Supplementary Table 6 The results of the likelihood ratio test when FPG and HbA1c levels were included in the model as continuous variables per 1-SD increase.

Models	Δ (vs. Model N) (Degrees of freedom)
		Δ (vs. each model)	Degrees of freedom	p-value
Model A FPG (continuous variables)	1.771 (1)	1.771a	1	0.183
Model B HbA1c (continuous variables)	1.455 (1)	1.455a	1	0.228
Model C FPG (continuous variables) + HbA1c (continuous variables)	2.884 (2)	1.113b	1	0.290
		1.429c	1	0.232

Δ = –2Log Likelihood (model k) – [–2Log Likelihood (model k + α)]

^a Relative to Model N.

^b Relative to Model A.

^c Relative to Model B.

Model N was a multiple logistic regression model with maternal age, pre-pregnancy BMI, gestational weight gain, hemoglobin, HDP in prior pregnancy, family history of hypertension, family history of diabetes mellitus, smoking status during pregnancy, alcohol intake, parity (primipara or not), ART, season at the time of blood collection, and initial MAP.

Model A was a multiple logistic regression model with FPG (per 1-SD increase), in addition to Model N.

Model B was a multiple logistic regression model including HbA1c (per 1-SD increase) in addition to Model N.

Model C was a multiple logistic regression model including both HbA1c (per 1-SD increase) and FPG (per 1-SD increase), in addition to Model N.

ART, assisted reproductive technology; BMI, body mass index; DBP, diastolic blood pressure; FPG, fasting plasma glucose; HbA1c, glycosylated hemoglobin; HDP, hypertensive disorders of pregnancy; MAP, mean arterial pressure

Supplementary Table 7 The area under the receiver operating characteristics curves (AUCs) when FPG or HbA1c levels at <24 weeks of gestation as continuous variables were included in a binary logistic regression model, for prediction of HDP.

Variables	AUC (95% CI)a	Cut-off value based on Youden index	Sensitivity (95% CI)	Specificity (95% CI)	PPV (95% CI)	NPV (95% CI)	Positive LR (95% CI)	Negative LR (95% CI)
FPG (mg/dL)	0.58 (0.52–0.63)	84	0.56 (0.48–0.64)	0.60 (0.57–0.63)	0.15 (0.12–0.19)	0.91 (0.90–0.93)	1.38 (1.17–1.63)	0.74 (0.61–0.90)
HbA1c (%)	0.59 (0.53–0.64)	5.2	0.40 (0.32–0.48)	0.78 (0.75–0.80)	0.19 (0.14–0.23)	0.91 (0.89–0.93)	1.77 (1.40–2.24)	0.78 (0.68–0.90)

^a Difference in the AUCs between FPG level and HbA1c level at <24 weeks of gestation was not statistically significant (p-value = 0.89 based on the Delong’s test)

Abbreviations: AUC, area under the receiver operating characteristics curve; CI, confidence interval; FPG, fasting plasma glucose; LR, likelihood ratio; NPV, negative predictive value; PPV, positive predictive value

References

• 1   Mol  BWJ,  Roberts  CT,  Thangaratinam  S,  Magee  LA,  De Groot  CJM, et al. (2016) Pre-eclampsia. Lancet 387: 999–1011.
• 2   (2020) Gestational hypertension and preeclampsia: ACOG practice bulletin, number 222. Obstet Gynecol 135: e237–e260.
• 3  HAPO Study Cooperative Research Group;  Metzger  BE,  Lowe  LP,  Dyer  AR,  Trimble  ER,  Chaovarindr  U, et al. (2008) Hyperglycemia and adverse pregnancy outcomes. N Engl J Med 358: 1991–2002.
• 4   Lowe  LP,  Metzger  BE,  Dyer  AR,  Lowe  J,  McCance  DR, et al. (2012) Hyperglycemia and adverse pregnancy outcome (HAPO) study: associations of maternal A1C and glucose with pregnancy outcomes. Diabetes Care 35: 574–580.
• 5   McIntyre  HD,  Sacks  DA,  Barbour  LA,  Feig  DS,  Catalano  PM, et al. (2016) Issues with the diagnosis and classification of hyperglycemia in early pregnancy. Diabetes Care 39: 53–54.
• 6   Benhalima  K,  Mathieu  C,  Damm  P,  Van Assche  A,  Devlieger  R, et al. (2015) A proposal for the use of uniform diagnostic criteria for gestational diabetes in Europe: an opinion paper by the European Board & College of Obstetrics and Gynaecology (EBCOG). Diabetologia 58: 1422–1429.
• 7   Iwama  N,  Sugiyama  T,  Metoki  H,  Kusaka  H,  Yaegashi  N, et al. (2019) Difference in the prevalence of gestational diabetes mellitus according to gestational age at 75-g oral glucose tolerance test in Japan: The Japan Assessment of Gestational Diabetes Mellitus Screening trial. J Diabetes Investig 10: 1576–1585.
• 8   Colagiuri  S,  Falavigna  M,  Agarwal  MM,  Boulvain  M,  Coetzee  E, et al. (2014) Strategies for implementing the WHO diagnostic criteria and classification of hyperglycaemia first detected in pregnancy. Diabetes Res Clin Pract 103: 364–372.
• 9   Minakami  H,  Maeda  T,  Fujii  T,  Hamada  H,  Iitsuka  Y, et al. (2014) Guidelines for obstetrical practice in Japan: Japan Society of Obstetrics and Gynecology (JSOG) and Japan Association of Obstetricians and Gynecologists (JAOG) 2014 edition. J Obstet Gynaecol Res 40: 1469–1499.
• 10  International Association of Diabetes and Pregnancy Study Groups Consensus Panel;  Metzger  BE,  Gabbe  SG,  Persson  B,  Buchanan  TA,  Catalano  PA, et al. (2010) International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diabetes Care 33: 676–682.
• 11   Hughes  RCE,  Moore  MP,  Gullam  JE,  Mohamed  K,  Rowan  J (2014) An early pregnancy HbA1c ≥5.9% (41 mmol/mol) is optimal for detecting diabetes and identifies women at increased risk of adverse pregnancy outcomes. Diabetes Care 37: 2953–2959.
• 12   Mane  L,  Flores-Le  Roux JA,  Benaiges  D,  Rodriguez  M,  Marcelo  I, et al. (2017) Role of first-trimester HbA1c as a predictor of adverse obstetric outcomes in a multiethnic cohort. J Clin Endocrinol Metab 102: 390–397.
• 13   Iwama  N,  Sugiyama  T,  Metoki  H,  Saito  M,  Hoshiai  T, et al. (2020) Associations between glycosylated hemoglobin level at less than 24 weeks of gestation and adverse pregnancy outcomes in Japan: The Japan Environment and Children’s Study (JECS). Diabetes Res Clin Pract 169: 108377.
• 14   Kayemba-Kay’s  S,  Peters  C,  Geary  MPP,  Hill  NR,  Mathews  DR, et al. (2013) Maternal hyperinsulinism and glycaemic status in the first trimester of pregnancy are associated with the development of pregnancy-induced hypertension and gestational diabetes. Eur J Endocrinol 168: 413–418.
• 15   Mañé  L,  Flores-Le  Roux JA,  Pedro-Botet  J,  Gortazar  L,  Chillarón  JJ, et al. (2019) Is fasting plasma glucose in early pregnancy a better predictor of adverse obstetric outcomes than glycated haemoglobin? Eur J Obstet Gynecol Reprod Biol 234: 79–84.
• 16   Sesmilo  G,  Prats  P,  Garcia  S,  Rodríguez  I,  Rodríguez-Melcón  A, et al. (2020) First-trimester fasting glycemia as a predictor of gestational diabetes (GDM) and adverse pregnancy outcomes. Acta Diabetol 57: 697–703.
• 17   (2018) ACOG committee opinion No. 743: low-dose aspirin use during pregnancy. Obstet Gynecol 132: e44–e52.
• 18   Chaemsaithong  P,  Pooh  RK,  Zheng  M,  Ma  R,  Chaiyasit  N, et al. (2019) Prospective evaluation of screening performance of first-trimester prediction models for preterm preeclampsia in an Asian population. Am J Obstet Gynecol 221: 650.e1–650.e16.
• 19   Metoki  H,  Ohkubo  T,  Watanabe  Y,  Nishimura  M,  Sato  Y, et al. (2008) Seasonal trends of blood pressure during pregnancy in Japan: the babies and their parents longitudinal observation in Suzuki memorial hospital in intrauterine period study. J Hypertens 26: 2406–2413.
• 20   Ishikuro  M,  Obara  T,  Metoki  H,  Ohkubo  T,  Yamamoto  M, et al. (2013) Blood pressure measured in the clinic and at home during pregnancy among nulliparous and multiparous women: The BOSHI study. Am J Hypertens 26: 141–148.
• 21   Kashiwagi  A,  Kasuga  M,  Araki  E,  Oka  Y,  Hanafusa  T, et al. (2012) International clinical harmonization of glycated hemoglobin in Japan: from Japan Diabetes Society to national glycohemoglobin standardization program values. J Diabetes Investig 3: 39–40.
• 22   Hoelzel  W,  Weykamp  C,  Jeppsson  JO,  Miedema  K,  Barr  JR, et al. (2004) IFCC reference system for measurement of hemoglobin A1c in human blood and the national standardization schemes in the United States, Japan, and Sweden: a method-comparison study. Clin Chem 50: 166–174.
• 23   Watanabe  K,  Naruse  K,  Tanaka  K,  Metoki  H,  Suzuki  Y (2013) Outline of definition and classification of “Pregnancy induced Hypertension (PIH)”. Hypertens Res Pregnancy 1: 3–4.
• 24   Umesawa  M,  Kobashi  G (2017) Epidemiology of hypertensive disorders in pregnancy: Prevalence, risk factors, predictors and prognosis. Hypertens Res 40: 213–220.
• 25   Shen  EX,  Moses  RG,  Oats  JJN,  Lowe  J,  McIntyre  HD (2019) Seasonality, temperature and pregnancy oral glucose tolerance test results in Australia. BMC Pregnancy Childbirth 19: 263.
• 26   Hedderson  MM,  Ferrara  A (2008) High blood pressure before and during early pregnancy is associated with an increased risk of gestational diabetes mellitus. Diabetes Care 31: 2362–2367.
• 27   Cnossen  JS,  Vollebregt  KC,  De Vrieze  N,  Ter  Riet G,  Mol  BWJ, et al. (2008) Accuracy of mean arterial pressure and blood pressure measurements in predicting pre-eclampsia: systematic review and meta-analysis. BMJ 336: 1117–1120.
• 28   Youden  WJ (1950) Index for rating diagnostic tests. Cancer 3: 32–35.
• 29   Allison  PD (2001) Missing data. Sage publications, Thousand Oaks, USA: 67–68.
• 30   Ho-Pham  LT,  Nguyen  UDT,  Tran  TX,  Nguyen  TV (2017) Discordance in the diagnosis of diabetes: comparison between HbA1c and fasting plasma glucose. PLoS One 12: e0182192.
• 31   Phelps  RL,  Metzger  BE,  Freinkel  N (1981) Carbohydrate metabolism in pregnancy. XVII. Diurnal profiles of plasma glucose, insulin, free fatty acids, triglycerides, cholesterol, and individual amino acids in late normal pregnancy. Am J Obstet Gynecol 140: 730–736.
• 32   Roberts  JM,  Hubel  CA (2009) The two stage model of preeclampsia: variations on the theme. Placenta 30 (Suppl A): S32–S37.
• 33   Cawyer  CR,  Horvat  D,  Leonard  D,  Allen  SR,  Jones  RO, et al. (2014) Hyperglycemia impairs cytotrophoblast function via stress signaling. Am J Obstet Gynecol 211: 541.e1–541.e8.
• 34   Seki  H (2014) Balance of antiangiogenic and angiogenic factors in the context of the etiology of preeclampsia. Acta Obstet Gynecol Scand 93: 959–964.
• 35   Nakanishi  S,  Aoki  S,  Kasai  J,  Shindo  R,  Obata  S, et al. (2020) High probability of false-positive gestational diabetes mellitus diagnosis during early pregnancy. BMJ Open Diabetes Res Care 8: e001234.
• 36   Cosson  E,  Vicaut  E,  Sandre-Banon  D,  Gary  F,  Pharisien  I, et al. (2019) Early screening for gestational diabetes mellitus is not associated with improved pregnancy outcomes: an observational study including 9795 women. Diabetes Metab 45: 465–472.
• 37   Brown  MA,  Magee  LA,  Kenny  LC,  Karumanchi  SA,  McCarthy  FP, et al. (2018) Hypertensive disorders of pregnancy: ISSHP classification, diagnosis, and management recommendations for international practice. Hypertension 72: 24–43.