IR harms the prognosis of GDM women
From the results of this retrospective study, IR was found to be harmful to the GDM’s prognosis, with an increase in adverse outcomes. When the confounding factors were excluded, the incidences of hypertensive disorders of pregnancy and LGA were significantly higher.
GDM with IR and hypertensive disorders of pregnancy
In our study, compared with the GDM women without IR, the risk of hypertensive disorders of pregnancy increased up to 5 times in GDM women with IR. An Iranian study  had similar findings: insulin resistance, closely related to preeclampsia, served as a risk factor in the process. Indeed, before clinical symptoms occur, the fasting insulin level increases accordingly as the disease progresses. Based on this, identifying GDM women with IR could be considered as a starting point for the clinical prevention of hypertensive disorders of pregnancy. However, the specific mechanism is still unclear. In the mainstream opinion, physiological IR provides more glucose to the fetus to better nourish its development, but overactivation of IR in pregnancy leads to decreased carbon monoxide (NO), discorded lipid metabolism, low prostaglandin E2 synthesis, and damaged vascular endothelial cells, eventually resulting in maternal high blood pressure [24, 25].
GDM with IR and LGA
GDM with IR may also cause fetal overgrowth. In our study, babies born to GDM women with IR had a significant increase in average birth weight compared with the babies of GDM women without IR. With the confounding factors (age, BMI, TG, FPG, HbA1c, etc.) adjusted, IR was found to be an independent risk factor of LGA, doubling the risk. This discovery is similar to that of Tressa Ellett’s research , in which IR was positively correlated to the birth weights of newborns. The mechanism of this phenomenon has been partially explained. In Pedersen's  hypothesis, maternal hyperglycemia passing through the placental barrier stimulates the fetus to secrete more insulin which promotes fat and protein accumulation, leading to excessive fetal growth. As a result, high-level maternal insulin, which though cannot go through the placenta, causes changes in the placenta’s metabolism, gene expression and epigenetic modifications , eventually disturbing fetal growth.
Glycaemic control/excessive GWG and adverse pregnancy outcomes
Our findings were inconsistent with the results of previous studies [29, 30]. In our study, poor glycemic control and excessive GWG increased the risk of hypertensive disorders of pregnancy, macrosomia, LGA and cesarean by univariable logistic regression analysis, but not by multivariable logistic regression analysis. Analysing the reasons of the abovementioned result, it was possible to consider that the sample size in certain subgroups was relatively small, so it leaded to a larger span of 95% CI. In addition, the inconsistent diagnosis criteria was another possible reason, that we employed criteria of pre-pregnancy BMI category  and excessive GWG  which were more suitable for Chinese people.
IR’s risk factors
Previous research has mainly focused on the risk factors of IR or GDM, but few referred to the case of GDM with IR, which we analyzed in our study. Based on our data, TG, HbA1c, and FPG in early pregnancy, pre-pregnant BMI, weight gain before diagnosis of GDM were considered to be the risk factors of GDM with IR. Our results are similar to the results of Sun’s study , in which BMI and TG were related to GDM with IR in the second trimester, but FPG and HbA1c were not. The differences in the results possibly derive from the fact that our FPG and HbA1c were tested in early pregnancy, while Sun’s results were obtained from the second trimester. Meanwhile, it is worth noting that FPG and HbA1c, although within normal range, were higher in GDM women with IR than that in those without IR, which indicates that more benefit will occur if blood sugar control is performed early. However, considering the dangers of hypoglycemia, more research needs to be done to determine the appropriate control range.
In addition, we found that weight gain before diagnosis of GDM was another risk factor of GDM with IR in the second trimester, which was consistent with literature reports. A retrospective study was conducted among 2,647 women with GDM which showed a significantly increased risk of IR for women with more weight gain before diagnosis of GDM . Thus, in future clinical work, we should enhance weight management initiating from the first trimester for high-risk women to improve insulin resistance.
Furthermore, in our research, age was a weak protective factor for IR. However, Sun reached the opposite conclusion . We consider that this difference in results comes from the fact that participants from Sun’ study were divided into four groups according to the severity of insulin resistance, defined by quartiles of HOMA-IR (1.540, 2.269, and 3.238), however, divided into two groups by HOMA-IR (2.0) for us. Therefore, uniform grouping criteria and expanding the study population are needed to confirm the results.
Management of GDM with IR
In this study, it was found that the HOMA-β value is increased in GDM women with IR. which is considered that IR promotes pancreatic β-cells to produce and secrete more insulin. This compensatory hyperinsulinemic response of β-cells can initially control blood glucose within the normal range; however, long-term exposure to excessive glucose and lipids will eventually lead to β-cell dysfunction and / or cell death, leading to overt diabetes. The protection of pancreatic β-cell function is closely related to the reduction of endogenous insulin demand . Therefore, women with insulin resistance to GDM are likely to benefit from insulin sensitization therapy, providing a new approach for the individualized treatment of GDM. This view is similar to that of a study in Poland, in which Sokup  suggested that GDM women with HOMA > 1.29 should be treated with improved tissue sensitivity to insulin, while GDM women with HOMA > 2.89 should be treated with metformin combined with insulin.
It is worth noting that the current clinical management goal of GDM is blood glucose. In fact, a high blood glucose level can only explain a small part of the variation in birth weight . In this study, although the most of women with GDM had a normal blood glucose range, women with GDM with insulin resistance had a higher risk of LGA. Our conclusion is consist with Li’s study . Therefore, insulin resistance should also be an important indicator in pregnancy monitoring, especially for GDM women within the normal range of blood glucose and with high risk factors.
In particular, in this study, only 1.7% of the GDM women were treated with insulin therapy to control blood glucose, which is quite different from in other countries. In Australia, this proportion can reach to 36.8% . This difference is considered to result from the lack of consensus and consistency in the screening and diagnostic criteria for gestational diabetes in various countries. In China, the diagnosis of GDM is up to the more rigid IADPSG standard, in which the cut-off value is lower than in other guidelines. The IADPSG criteria increase the prevalence of GDM . However, alternative standards are still used in some centers and countries including Australia . This has led to the expansion of the GDM population in China, while the proportion of women who really need insulin treatment has decreased. Therefore, it is worth exploring whether the IADPSG standard is applicable to the Asian population. We further suggest an appropriate stratification of the management of GDM women based on HOMA-IR to reduce the loss of medical resources. In addition, there are still 18 women in this study with poor glycemic control, due to various reasons (because of the retrospective study, the exact reason cannot be obtained), did not use insulin therapy, which was also one of the reasons for the low rate of insulin use in this study.
First, this study was a retrospective study, in which bias was inevitably introduced in the participant selection. Second, our research subjects were GDM pregnant women, so the results would not be applicable to non-GDM women. Third, our model did not include other risk factors that affect IR, including diet, physical activity, smoking and drinking. In addition, HOMA-IR before childbirth was not tested, so it was not clear whether the women’s IR was improved by lifestyle and diet management. In the end, more research is needed to dynamically observe the changes in IR before pregnancy, during pregnancy, and after delivery, and to clarify the influencing factors of IR and the relationship between IR and GDM.