Gestational diabetes mellitus (GDM), as the most common metabolic complication of pregnancy, poses a serious threat to maternal and fetal metabolic health. However, current GDM diagnosis faces several problems such as static, single-point, cumbersome to operate and delayed diagnosis, highlighting an urgent need to establish an individualized system for early prediction, diagnosis, and intervention.

This project aims to develop a mother-child cohort covering pregnancy and the perinatal period to propose early diagnostic criteria for GDM based on continuous glucose monitoring (CGM) technology, as well as developing clinically applicable AI-based tools for analyzing and interpreting CGM data, along with strategies to assist in GDM diagnosis. Furthermore, it will investigate CGM parameters and multi-omics biomarkers suitable for predicting maternal and fetal outcomes, culminating in the creation of an intelligent management platform for GDM. This project is expected to enhance the early identification rate of gestational diabetes, potentially advancing the diagnostic and therapeutic window for the condition, thereby improving both short- and long-term maternal and fetal health outcomes.

China has a diabetic population of 233 million, posing an enormous societal burden. Early recognition and intervention for diabetes are urgently needed. In recent years, growing evidence has highlighted the critical role of the early-life developmental environment in the pathogenesis of diabetes. As early as 1986, Professor Barker proposed the Developmental Origins of Health and Disease (DOHaD) theory. The investigators previously validated this theory for the first time in the Chinese population through cohort studies, demonstrating that adverse intrauterine environments lead to abnormal glucose metabolism in offspring and that early-life interventions can effectively prevent adult-onset diabetes.

According to the International Diabetes Federation's Diabetes Atlas (11th edition), the global incidence of hyperglycemia during pregnancy is 16.7%, with gestational diabetes mellitus (GDM) accounting for up to 80% of cases. This means that one in five live births is exposed to an adverse intrauterine environment early in life, increasing their risk of metabolic disorders such as overweight, obesity, and diabetes in adulthood. GDM significantly raises the risk of adverse pregnancy outcomes and seriously threatens the metabolic health of both mothers and offspring. Early and efficient diagnosis and prevention of GDM are therefore crucial for improving metabolic health in mothers and children.

The current diagnosis of GDM relies on oral glucose tolerance tests (OGTT) performed at 24-28 weeks of gestation, which present limitations such as static and single-time-point measurement, operational complexity, delayed diagnosis, and limited time for effective intervention. Thus, there is an urgent need to develop novel technologies for early prediction and diagnosis of GDM.

Continuous glucose monitoring (CGM) in the first trimester offers advantages including 24/7 detailed glucose data, detection of hidden hyperglycemia, assessment of glycemic variability, and compatibility with AI-assisted analysis, showing great potential for early diagnosis and management of GDM. Previously, the investigators applied CGM in patients with type 2 diabetes and was granted a Chinese invention patent for "Using CGM for Improved Management and Monitoring of Glucose in Type 2 Diabetes (CN 109637677A)." In recent years, CGM has been widely used in diabetes management and has begun to be applied in managing HbA1c levels in pregnant women with type 1 diabetes. However, research on its use in pregnant women with type 2 diabetes and GDM is still in its early stages. There is currently a lack of studies utilizing CGM combined with artificial intelligence for early diagnosis and prediction of GDM in the first trimester.

Besides, multiple studies have explored risk factors and biomarkers for GDM to enable early screening and predict maternal and fetal outcomes. However, most research has been limited to single-omics approaches or later gestational time points, presenting numerous constraints. Studies conducted at earlier gestational periods, across multiple time points, and utilizing multi-omics approaches will further reveal biomarkers predictive of maternal and fetal outcomes in GDM.

Therefore, the investigators plan to establish a GDM mother-child cohort covering the pregnancy and perinatal periods. They aim to propose early diagnostic criteria for GDM based on CGM, develop clinically applicable AI-driven tools for analyzing and interpreting CGM data, and formulate auxiliary diagnostic strategies for GDM. The investigators will explore CGM parameters and multi-omics biomarkers suitable for predicting maternal and fetal outcomes. Based on these findings, the investigators are expected to establish an intelligent management platform for GDM, which will move forward the clinical window for GDM diagnosis and treatment, improving both short- and long-term health outcomes for mothers and their offspring.

Inclusion Criteria:

• ① Early pregnancy (≤14 weeks) pregnant women;
• ② Singleton pregnancies;
• ③ Early pregnancy psychological scores (PHQ-9 and GAD-7) <10 points;
• ④ Consent to participate in the study and sign an informed consent form.

Exclusion Criteria:

• ① Twin or multiple pregnancies;
• ② Diabetes mellitus complicated with pregnancy;
• ③ Severe pregnancy complications;
• ④ Pre-existing significant cardiovascular, hepatic, renal, hematologic, or autoimmune diseases;
• ⑤ History of smoking, alcohol abuse, or narcotic and drug use;
• ⑥ Early pregnancy psychological assessment (PHQ-9 or GAD-7) score ≥10.

women with gestational diabetes and control group