The relationship between poor glycaemic control at different time points of gestational diabetes mellitus and pregnancy outcomes

Abstract We aimed to identify the complications of gestational diabetes mellitus (GDM) associated with poor control of fasting plasma glucose (FPG) and postload plasma glucose (PPG) on the 75-g oral glucose tolerance test (OGTT). This retrospective study included 997 singleton pregnancy GDM patients who were assigned to poor or good glycaemic control groups. Multivariate analysis indicated that poor FPG control and poor PPG control were both independent predictors of hypertensive disorder complicating pregnancy (HDCP) (odd ratio (OR) of 2.551 (95% CI [1.146–5.682], p = .022) and OR of 2.084 (95% [1.115–3.894], p = .021) compared with good glycaemic control groups, respectively). Poor PPG control promoted the rate of caesarean delivery (1.534 (95% CI [1.063–2.214]), p = .022), whereas good PPG control increased the risk of premature rupture of membranes (PROM) (0.373 (95% CI [0.228–0.611]), p < .001). Conclusively, poor control FPG and PPG dissimilarly affect pregnancy complications in GDM; these findings may help clinicians in the effective implementation of measures to prevent pregnancy complications in GDM. IMPACT STATEMENT What is already known on this subject? Previous studies displayed that GDM patients with 2-h PPG elevated at 24–28 week of gestation had a 2.254-fold increased risk of postpartum dysglycaemia. Abnormal plasma glucose in GDM mother increased the probability of childhood obesity in the offspring. With the implementation of China’s second-child policy, the incidence of GDM is rising. What do the results of this study add? Our results indicated that the older patients with GDM, the greater the risk of abnormal plasma glucose control. In addition, maternal age and prenatal BMI were notably correlated with poor plasma glucose control of FPG and PPG, respectively. We also found that both poor FPG and PPG control notably increased the incidence of HDCP in pregnant women. The incidence of PROM was higher in the good PPG control group compared with the poor PPG control group. What are the implications of these findings for clinical practice and/or further research? This study displayed that the effects of poor FPG and PPG control on pregnancy complications and newborn outcomes were heterogeneous, which might be related to the specificity of plasma glucose metabolism at different time points. Good glycaemic control, especially PPG control, was of great significance for improving pregnancy complications and perinatal conditions.


Introduction
Gestational diabetes mellitus (GDM) is a singular form of diabetes, which is mainly characterised by glucose metabolism disorders during pregnancy (Leng et al. 2015). In recent years, with the rapid increase in obesity and diabetes prevalence, the global incidence of GDM is also increasing; the global prevalence of GDM was estimated to be 12.8% by the 2019 version of Diabetes Atlas Yuen et al. 2019). Clinical data showed that GDM mostly occurs in the middle and late stages of pregnancy, and is more common in elderly women (Ferrara 2007;Yuen et al. 2019). The negative effects of GDM include increased risk for hypertensive disorder complicating pregnancy (HDCP), caesarean section, induction of labour, large for gestational age, shoulder dystocia, foetal macrosomia and neonatal asphyxia (Farrar et al. 2016). Current evidences indicate that GDM patient have a greater risk of developing type 2 diabetes (Vounzoulaki et al. 2020), other metabolic complications (Mitanchez 2010) and cardiovascular diseases (Harreiter et al. 2014) in the future. GDM is a high-risk pregnancy disease, which not only affects the process of pregnancy and childbirth, but also lays down hidden dangers for the future health of pregnant women and offspring (Hillier et al. 2007;Wang et al. 2019).
In 1998, the World Health Organisation recommended diagnostic criteria for diabetes, but did not distinguish between pregnant and non-pregnant women (Alberti et al. 1998). In 2010, the International Association of Diabetes and Pregnancy Study Groups (IADPSG) published new diagnostic standards based on the study of Hyperglycaemia and Adverse Pregnancy Outcomes (HAPO) (Hadar and Hod 2010). The diagnostic criteria for pregnant women have now been identified, and attention has been drawn to the influence of abnormal plasma glucose on adverse pregnancy outcomes. Currently in China, following the IADPSG standard, an oral glucose tolerance test (OGTT) is done during the 24 th to 28 th week of pregnancy. Fasting plasma glucose (FPG) and 1 hand 2-h plasma glucose on OGTT are measured. If the result meets or exceeds the standard of FPG ranged between 5.1 and 6.9 mmol/L, 1-h plasma glucose �10.0 mmol/L or 2-h plasma glucose in the range of 8.5-11.0 mmol/L, respectively, a diagnosis of GDM is confirmed (Diagnostic criteria and classification of hyperglycaemia first detected in pregnancy: a World Health Organisation Guideline 2014). FPG reflects whether the secretory function of pancreatic b-cells can control the output of liver glycogen, which is an indicator of glucose metabolism in a resting or non-feeding state (R€ oder et al. 2016). Several recent studies have incriminated elevated FPG as a high risk factor for a variety of diseases, for instance, pancreatic cancer (Nagai et al. 2017) and cerebral infarction (Cao et al. 2015). In addition, it has been reported that among non-pregnant women, elevated plasma glucose levels, especially 2 h after meals, were associated with fatal and non-fatal cardiovascular events (Shahim et al. 2017). A previous study displayed that GDM patients with 2 h PPG elevated at 24-28 week of gestation had a 2.254-fold increased risk of postpartum dysglycaemia (Wang et al. 2019). Moreover, Hillier et al. (2007) found that abnormal plasma glucose in GDM mother increased the probability of childhood obesity in the offspring. Also, a previous study displayed that optimising PPG control improves cardiovascular-related diseases (Bibra et al. 2009). With the implementation of China's second-child policy, the incidence of GDM is rising and the increasing trend of GDM has attracted continuous attention in the field of obstetrics. However, the current state of the management of Chinese GDM patients is not optimal, which needs improvement.
Glycaemic control is valuable in the handling of GDM patients and for preventing pregnancy complications and outcomes. Nevertheless, research on the correlation between glycaemic control and pregnancy outcomes, maternal health and newborn health is limited (Yefet et al. 2019). Additional studies may be useful in minimising the burden of GDM via the facilitation of the keeping of blood glucose levels of GDM patients within the normal range. Therefore, identifying the complications associated with poor glycaemic control of GDM patients, and gaging the impact of glycaemic control on maternal and foetal outcomes are essential for guiding the prevention of GDM.
Thus, this study was designed to reveal the adverse outcomes of GDM that are linked with poor glycaemic control on the basis of FPG and PPG values on 75-g OGTT and the discrepancies between the impact of poor FPG control and that of poor PPG control.

Study population
Among the 9666 patients who were hospitalised for delivery in Changzhou Second People's Hospital affiliated to Nanjing Medical University from 1 January 2016 to 31 December 2017, 1569 (16.23%) were diagnosed with GDM. The 380 cases for whom the 75-g OGTT was not performed and who were only diagnosed with elevated fasting blood glucose were excluded; 23 cases of chronic hypertension complicated with pregnancy and 24 cases of twin pregnancies were excluded; 145 cases of patients with endocrine diseases before pregnancy were excluded. Finally, the study included 997 singleton pregnancy GDM patients. This retrospective study was approved by the Institutional Review Boards of the affiliated Changzhou 2nd People's Hospital of Nanjing Medical University. All procedures implicating human individuals were in accordance with the ethical standards of the Institutional Review Boards of the affiliated Changzhou Second People's Hospital of Nanjing Medical University and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Inclusion criteria
The included GDM patients met the following inclusion criteria: (1) women with singleton pregnancy; (2) 75-g OGTT was done at 24-30 weeks; (3) no endocrine and metabolic diseases such as diabetes or hyperthyroidism before pregnancy; (4) No hypertension, coronary heart disease and other cardiovascular diseases before pregnancy; (5) no neurological diseases; (6) patients with non-smoking and non-alcohol abuse and other bad habits; (7) gestational age between 28 and 42 weeks and (8) available and complete case data.

Exclusion criteria
Women with twin or multiple pregnancies, pre-pregnancy diabetes, severe polycystic ovary syndrome, immune system disease, scarred uterus, mental illness and unavailable and incomplete case data were excluded from the study.

OGTT method and diagnosis of GDM
All participants underwent a 75-g OGTT in the 24 th to 28 th week of gestation. Briefly, 75 g of glucose water was administered orally within 5 min, and then intravenous FPG, 1-h and 2-h postload plasma glucose (PPG) values based on OGTT were measured. The time was gaged from the administration time of oral glucose water, and the plasma glucose value was obtained via enzymatic approaches with the Hitachi Automatic Analyser 7600 (Hitachi, Tokyo, Japan). The upper limits of normal for FPG, 1-h PPG, and 2-h PPG were 5.1 mmol/h, 10.0 mmol/h and 8.5 mmol/h, respectively. GDM was diagnosed when one item reached or exceeded the critical value. After diagnosis, glycaemic control of patients was initiated during their first specialised appointment to achieve their glycaemic targets.

Definition of research related indicators
FPG was defined as the blood glucose detected from venous blood following at least 8 h of overnight fasting (Marathe et al. 2017). Glycaemic control is the technique of supervising the blood glucose level of diabetic patients at optimum level (Marathe et al. 2017). The target values of glycaemic control were 3.3-5.3 mmol/L for FPG and 4.4-6.7 mmol/L for 2-h PPG. If the glycaemic control does not reach the target value after diet, exercise or after the use of insulin, it is considered that the glycaemic control is poor. Good glycaemic control was defined if the patients had 4 and 7 mmol/L of FPG upon measurement of three consecutive visits. The observed adverse pregnancy outcomes were: HDCP, PROM, postpartum haemorrhage (PPH), macrosomia, low-birth weight infants, premature delivery, neonatal jaundice, neonatal asphyxia and indicators of insulin use; male newborns and caesarean section were also observed. The general information of each group included age, pregnancy times, parity, antenatal BMI value and new born weight (NBW). Adverse pregnancy outcome was declared by the occurrence of HDCP, PROM, PPH and macrosomia, premature birth, low birth weight, neonatal asphyxia and neonatal jaundice, otherwise there was no adverse pregnancy outcome. Prenatal BMI refers to the body mass index before delivery, which refers to weight divided by the square of height (kg/m 2 ).

Study design
Patients with abnormal glucose sugar levels were referred to the nutrition clinic or diabetes clinic for treatment according to the standard of care, where they were informed about diet and exercise therapeutic methods and given target values of glycaemic control (FPG concentration < 95 mg/dL and 2-h PPG < 120 mg/dL) (Schwartz et al. 2018;Yefet et al. 2019). The diet for overweight and obese women was 25 kcal/kg, whereas the diet of patients with normal weight was 35 kcal/kg. The diets were given as full meals (three per day) and four snacks containing 30% fat, 50% carbohydrates and 20% protein. Self-monitoring of plasma glucose of patients was achieved using a daily chart including three preprandial measurements, three post-prandial measurements 2 h after meals and a single measurement at before bed. The daily chart measurements lasted for a week. After this time, if 20% of glycaemic measurements of pre-prandial or postprandial glucose values were greater than 95 mg/dL > 120 mg/dL, respectively, insulin treatment was initiated. The daily glucose charts lasted until delivery. Prior to their appointment, scheduled women followed a controlled daily diet consisting of at least 200 g of carbohydrates over a three-day period. On the day of the appointment, women came after fasting overnight for OGTT following informed consent. Women's initial weights were measured at 8-10 week of gestation during the first prenatal visit. The same mechanical column scale was used for weighing patients with a minimum of clothing at each visit. A visit was achieved every four weeks. The last values of blood pressure and maternal weight at 34-36 weeks before delivery were obtained from hospital records. If the patient's glycaemic control value did not reach the target after two weeks, insulin medication was recommended. According to published standards, patients diagnosed with GDM who received diet, exercise or insulin therapy were defined as poor glycaemic control (mean self-monitoring daily plasma glucose more than 95 mg/ld.), otherwise good glycaemic control (Yefet et al. 2019). Of the 997 patients enrolled in the study, 476 had good glycaemic control and 521 had poor glycaemic control. According to the plasma glucose levels measured at different time points based on OGTT, GDM patients with poor glycaemic control were further divided into two groups: poor FPG control and poor PPG control (1-h PPG or/and 2-h PPG greater than the guideline).

Sample size calculation
The appropriate sample size was obtained with the online sample size calculator (https://www.surveysystem.com/sscalc. htm) considering a confidence level of 95% and an alpha level of 5%. As a result, the sample size was estimated to be N ¼ 278 for the population of 997 GDM patients.

Data collection
Information on baseline characteristics and pregnancy outcome were collected from the pregnant women's medical records, gestational diabetes clinical records, laboratory system storage data and delivery records. The follow-up period began with the GDM diagnosis and ended with the re-examination at six weeks postpartum. Maternal demographic and basic clinical data included age, height, gravidity, parity, prenatal weight, prenatal BMI and intervention (diet and exercise or insulin therapy). The following information was included in pregnancy complications: caesarean delivery, PPH, HDCP, PROM and scarred uterus. Neonatal health parameters include neonatal jaundice, neonatal asphyxia, birth weight (neonatal weight less than 2500 g is defined as low birth weight and neonatal weight over 4000 g is foetal macrosomia), premature delivery (delivery before 37 weeks of pregnancy) and adverse pregnancy outcomes. Birth weights of neonates were obtained from maternity records. The weights of the naked newborns were taken immediately after delivery using the electronic scale.

Statistical analyses
Numerical data were processed using SPSS 25.0 (IBM Corp., Armonk, NY, USA). Continuous variables were presented as means ± standard deviation (SD), whereas frequencies and proportions were used for expressing the categorical variables. For continuous variables, Levene test was employed for assessing the homogeneity of variance, whereas the Shapiro-Wilk test was used for gaging the normal distribution of the variables. The Mann-Whitney-U test (for non-normally distributed variables) or unpaired t-test (normally distributed variables) was used to analyse the discrepancies among the good glycaemic control and poor glycaemic control groups. The categorical variables were analysed using the Chisquared test. Binary logistic regression was applied to explore the effects of glycaemic control on pregnancy complications and newborn outcomes. Parameters of women's baseline characteristics were adjusted in multivariate logistic regression analysis as followed: age (<20 y, 20-24 y, 25-29 y, �30 y); height (�160 cm, >160 cm); gravidity (�2, >2); prenatal weight (�70 kg, >70 kg); prenatal BMI (underweight: <18.5 kg/m 2 , normal: 18.5-24.9 kg/m 2 , overweight: 25.0-29.9 kg/m 2 , obese: �30.0 kg/m 2 ) (Darling et al. 2014). p < .05 was considered statistically remarkable.

Correlation of poor glycaemic control with demographic factors
As shown in the Tables 1 and 2, out of 997 patients included in our study, the majority of patients (46.84%) were aged 25-29 years (n ¼ 467). There were 768 (77.03%) patients whose prenatal BMI exceeded the normal range, including 500 (50.15%) overweight and 268 (26.88%) obese (Table S1). However, only 2.31% (n ¼ 11) of patients with good glycaemic control and 3.84% (n ¼ 20) of patients with poor glycaemic control chose insulin therapy (Table 1). To explore the risk factors for poor glycaemic control, univariate analysis was used to analyse the discrepancy of demographic factors between GDM patients with good and poor glycaemic control. It was found that maternal age was a predictor of glycaemic control when the time period was not considered, and older patients had a greater risk of poor glycaemic control (p ¼ .003 ; Table 1). Similarly, the age of pregnant women was pointedly greater in patients with poor FPG control (28.84 ± 4.07) than that in patients with good FPG control (27.77 ± 3.83, p ¼ .026; Table 1). In patients with poor PPG control, the prenatal BMI level was the only demographic factor that had a remarkable association with glycaemic control (p ¼ .013; Table 1).

The influence of poor glycaemic control on pregnancy outcomes
After adjusting the baseline characteristics of GDM patients between the good and poor glycaemic control groups, we used the Chi-squared test to analyse the different variables, including maternal age, height, gravidity, parity, prenatal weight, prenatal BMI and intervention. The results displayed that there was no remarkable discrepancy in demographic factors between the two groups (all p > .05, Table S1).
To investigate the correlations of glycaemic control during pregnancy with the pregnancy complications and newborn outcomes, univariate and multivariate models were developed to determine if poor glycaemic control impacted pregnancy outcomes. The results of univariate and multivariate analysis consistently displayed that the number of deliveries by caesarean section (multivariate OR, 1.426; 95% CI, 1.105-1.841; p ¼ .006), the incidence of HDCP (multivariate OR, 1.737; 95% CI, 1.182-2.552; p ¼ .005), the occurrence of scarred uterus (multivariate OR, 1.448; 95% CI, 1.033-2.028; p ¼ .031) and the number of foetal macrosomia (multivariate OR, 1.631; 95% CI, 1.134-2.346; p ¼ .008) in the poor glycaemic control group were notably greater than those in the good glycaemic control group. Unexpectedly, the incidence of PROM (multivariate OR, 0.589; 95% CI, 0.418-0.829; p ¼ .002) in the good glycaemic control group was found to be notably greater than that in the poor glycaemic control group. There was no remarkable discrepancy found between the two groups in instances of PPH, premature delivery, neonatal jaundice, neonatal asphyxia, low birth weight and adverse pregnancy outcomes (all p > .05, Table 2).

The influence of poor glycaemic control at different time points on pregnancy outcomes
To further explore the effects of poor glycaemic control at different time points on pregnancy outcomes, adjustments on the basic characteristics of GDM patients between the good and poor glycaemic control groups at different time points (FPG and PPG) were made. As shown in Table S2, no remarkable discrepancy in patient baseline characteristics, such as maternal age, height, gravidity, parity, prenatal weight, prenatal BMI, and intervention, was observed between the good and poor glycaemic control groups in either FPG or PPG (all p > .05, Table S2).
Similarly, we established univariate and multivariate models to evaluate the impacts of poor glycaemic control at different time points on pregnancy outcomes. Correlations of poor glycaemic control with the pregnancy complications and newborn outcomes are shown in Table 3. In both univariate and multivariate analyses, uncontrolled increase in FPG level notably elevated the risk of HDCP in pregnant women (multivariate OR, 2.551; 95% CI, 1.146-5.682; p ¼ .022), but poor FPG control did not notably promote other pregnancy outcome indicators, including caesarean delivery, PPH, PROM, scarred uterus, neonatal jaundice, neonatal asphyxia, foetal macrosomia, low birth weight, premature delivery and adverse pregnancy outcomes (all p > .05, Table 3). In addition, poor PPG control not only promoted the risk of HDCP in GDM patients (multivariate OR, 2.084; 95% CI, 1.115-3.894; p ¼ .021), but also notably increased the number of caesarean sections (multivariate OR, 1.534; 95% CI, 1.063-2.214; p ¼ .022, Table 3). In addition, the risk of PROM was notably increased in pregnant women with good PPG control compared with the patients with poor PPG (multivariate OR, 0.373; 95% CI, 0.228-0.611; p < .001, Tables 2 and 3).

Discussion
Poor maternal glycaemic control fosters the risk for adverse pregnancy outcomes, and stringent management of GDM is necessary to avoid maternal and neonatal complications. However, complications associated with poor glycaemic control and the impact on pregnancy and newborn outcomes are not well-investigated in the pregnant women with GDM. In this investigation, we explored the key predictors of glycaemic control during pregnancy and assessed the impacts of poor glycaemic control on pregnancy outcomes. We also examined the effect of FPG and PPG on the obstetrics and neonatal outcomes. We found that maternal age was a predictor of poor glycaemic control and older patients had a greater risk of poor glycaemic control. The age of patients with poor FPG control or poor PPG control was greater than that of good FPG control or good PPG control. The number of deliveries by caesarean section, the incidence of HDCP, the occurrence of scarred uterus and the number of foetal macrosomia were notably greater in the poor glycaemic control group compared with the good glycaemic control group. In addition, poor FPG control notably increased the risk of HDCP in pregnant women, whereas poor PPG control not only promoted the risk of HDCP in GDM patients but also notably increased the number of caesarean sections. In addition, poor PPG control was followed by a decrease in PROM risk. We concluded that the management of FPG and PPG may be crucial in avoiding serious pregnancy complications in women with GDM. This study identified key predictors associated with poor glycaemic control. First, the study displayed that the older the patients with GDM, the greater the severity of their glycaemic disorders. Regardless of whether the blood glucose is well controlled or not, the increase in PPG was notably correlated with age. This study displayed that the average age of severely ill patients was over 30 years old. In addition, maternal age and prenatal BMI were notably correlated with poor plasma glucose control of FPG and PPG, respectively. This was consistent with findings from Makgoba et al. (2012) that there is a strong positive correlation between the reproductive age and GDM. Some scholars have studied the mechanism of age on blood sugar, and displayed that as age increases, the sensitivity of the tissues to insulin decreases and leptin levels increase (Aarsand et al. 2018), which leads to increased blood glucose levels. But the above study did not explain whether FPG or PPG is closely related to age. Our study displayed that PPG and age of pregnant women were notably correlated. Although there are very few such studies during pregnancy, there is no consensus argument for large samples and multiple centres. However, in the non-pregnant population, the relationship between PPG and cardiovascular events, especially the relationship between PPG and fatal and non-fatal cardiovascular events has been paid attention. Combined with the results of this study, it is recommended to procreate at a suitable childbearing age to reduce the occurrence of GDM. The older the parturient is, the earlier intervention is needed to prevent the occurrence of GDM. This study also displayed that the prenatal BMI of the poorly controlled group was greater than that of the well-controlled group, suggesting that poor blood sugar control may be affected by weight gain during pregnancy. This result was similar to previous studies which revealed that as the prenatal BMI value increases, the risk of adverse pregnancy outcomes also increases (Simko et al. 2019;Sun et al. 2020). Few studies have indicated the association of HDCP with pregnant women in general and with GDM in particular (Yan et al. 2021). HDCP is known to be associated with low birth weight (Rao et al. 2018) and has been demonstrated as a risk predictor in GDM (Yao et al. 2018). The independent predictors of HDCP in GDM are not well elucidated and the related studies are very limited. It was previously reported that HDCP can be increased by pre-pregnancy overweight (Lin et al. 2019). Another study indicated that maternal lipid profiles in the first trimester of pregnancy are correlated with HDCP and GDM (Zhang et al. 2021). It was also reported that HDCP in GDM patients is correlated with the expression of microRNA-518 (Li et al. 2022). However, only one previous study showed that the increase of OGTT thresholds increases the risk of maternal morbidity in terms of HDCP (Savona-Ventura and Chircop 2008), suggesting that glycaemic control may be important for limiting the adverse effects associated with GDM. However, the relationship of FPG and PPG with HDCP has not been reported so far. The major finding of the present investigation was that poor FPG control and poor PPG control differentially affect the pregnancy and foetal outcomes in GDM and the common factor associated with both poor FPG control and poor PPG control was increased HDCP. The incidence of HDCP in patients with good glycaemic control was notably lower in this study, suggesting that it is not only necessary to control plasma glucose in pregnant women, but it is also important to focus on the prevention of HDCP in patients with GDM to prevent the adverse outcomes for the maternal and neonatal health. The increase of HDCP in patients with poor glycaemic control may be the result of abnormal glucose metabolism which has been demonstrated to induce hypertension (Corry and Tuck 1996).
PROM occurs in 5-10% of pregnancies and is responsible for an important proportion of neonatal mortality and morbidity (Yasmina and Barakat 2017). The factors associated with PROM are multifarious and only some of them have been reported, which necessitates continual effort in identifying these factors. PROM is known to be a complication associated with GDM; however, studies focussed on the identification of factors associated with PROM and possible ways to mitigate its adverse effects are limited. Up to date, the association of PROM with glycaemic control in GDM patients is unclear. Our study indicated that the risk of PROM was notably increased in pregnant women with good PPG control, suggesting that maintaining PPG in normal range instigates PROM in GDM patients. This result was contradictory to previous studies showing that PROM is one of the most common maternal complications in women with GDM and is associated with poor glycaemic control. Our study suggested that the good control of PPG may induce changes in some factors, which ultimately promotes PROM. These factors are probably those involved in glycolysis and fatty acid synthesis. Indeed, a previous study on metabolite changes in the vagina of women with PROM indicated that glycolysis was the main metabolic pathway that is downregulated in PROM (Liu et al. 2021). The downregulation of the upstream metabolites of glycolysis (N-acetyl-d-galactosamine and sucrose) signposted that, in PROM, the glycolysis pathway is not fully fuelled, which influences the capacity of the Lactobacillus species to maintain the normal pH in the vagina through the production of lactic acid (Liu et al. 2021). Another study indicated that the downregulation of proteins regulating glycolysis, which leads to an imbalance of glycolysis and reduced supply of pyruvate, triggers lactate oxidation in mitochondria (Pan et al. 2020). This activation of lactate oxidation generates pyruvate to compensate the shortage of pyruvate in the conventional glycolysis pathway. This leads to the decrease of energy supply in the foetal membrane because the glucose/lactate is the main source of energy for foetal development (Kane 2014;Pan et al. 2020). Since placenta is the principal source of lactate in gestation, the imbalance of glycolysis due to glycaemic control as well as the disruption of fatty acid metabolism in the maternal-placental interface may result in foetal malnutrition and preterm birth (Makinde et al. 1998;Bobi� nski and Mikulska 2015;Pan et al. 2020). Thus, we speculated that the good control of PPG may induce the imbalance of glycolysis in the foetal membrane, thus leading to a weakened membrane susceptible to PROM. Our study highlighted that though glycaemic control is vital in preventing some maternal and foetal outcomes, it may also be deleterious in other cases. It is vital to screen GDM patients to identify those who are fitted for glycaemic control; this may involve approaches such as metabolomics and transcriptomics instead of OGTT alone.
Patients with GDM are prone to excess amniotic fluid, which leads to high uterine cavity pressure, resulting in PROM; however, no patients with excessive amniotic fluid were detected in the population observed in the current study, and this must be due to the fact that GDM patients were mostly patients with milder disease and were controlled by diet plus exercise, resulting in minimal occurrence of amniotic fluid. However, a negative correlation between PROM and poor PPG was established in this investigation, indicating that the good control of PPG in OGTT may induce PROM in GDM patients. But still, a larger sample size and multicentre studies are required for further validation.
This study also displayed that the incidence of caesarean delivery increased notably with the severity of glucose abnormalities (high PPG) in GDM, suggesting that the severity of GDM leads to increased adverse pregnancy outcomes. This was consistent with the study of Papachatzopoulou et al. (2020). The results were also in corroboration with previous studies indicating that abnormal glucose values in OGTT are highly correlated with high risk of caesarean section (Bhavadharini et al. 2022). It was also reported that FPG is associated with caesarean section (Tennant et al. 2022). This also suggests that the more severe the disorder of glucose metabolism situation, the more it leads to maternal and child outcomes. These discrepancies indicated that a rigorous control of PPG is fundamental for unearthing the risk of and avoiding negative pregnancy complications in patients with GDM.
This study presents some limitations. First, it was a retrospective design with unavoidable selection bias. Second, it was a single-centre and this may restrict the worldwide application. Thus, further evaluation and studies are still necessary.
In conclusion, this study displayed that the effects of poor FPG and poor PPG control on pregnancy complications and newborn outcomes were heterogeneous, which might be related to the specificity of plasma glucose metabolism at different time points. Good glycaemic control, especially PPG control, is of great significance for improving pregnancy complications and perinatal conditions. Due to the inherent limitations of retrospective studies, larger scaled prospective studies are needed in the future to confirm these findings and provide a clinical basis for the stratified management model of pregnant women with GDM.

Ethical approval
Research involving human material and human data has been performed in accordance with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. All of our experiments were permitted by the Ethics Committee of The Affiliated Changzhou 2nd People's Hospital of Nanjing Medical University (Permission No. KY201226). As part of the ethics clearance, the hospital director consented to the use of the data. Patients consented to the analysis of their medical records. Written informed consent was signed by all individual participants to confirm their agreement to participate in the study.

Author contributions
ML designed the study and revised the manuscript. HX analysed data and wrote the first draft of manuscript. XT and JL conducted experiments and collected data. All authors read and approved the final version of manuscript.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This study was supported by the Fund of Changzhou Municipal Health Commission (QN201816) and Changzhou Sci&Tech Program (CJ20220137).