研究背景：妊娠期糖尿病（gestational diabetes mellitus，GDM）已被证实与孕产妇和新生儿的短期和长期不良结局及代谢异常风险增加密切相关，其早期识别和干预将显著改善母胎预后。筛查和鉴定新的母体血清风险预测生物标志物，依然是当前国际国内GDM疾病防控与临床早期干预研究的热点和挑战。然而，既往研究多数仅提供了针对单个或几个细胞因子或蛋白的研究证据，仅仅关注单一妊娠窗口，不同程度糖耐量受损的GDM亚型也未得到细分，且存在研究结果不一致的情况。因此，本研究从糖脂代谢网络通路的角度系统选择包括血清多肽标志物（C-peptide、GIP、PP）、炎性细胞因子（sCD40L）、脂肪细胞因子（Adiponectin、Leptin、Resistin、Adipsin）在内的8个细胞因子，评估不同血清生物标志物在妊娠女性妊娠早中期血清中的水平与GDM不同亚型风险的关联性，分别构建妊娠早期和妊娠中期的GDM疾病风险预测模型，为促进GDM高危人群早期筛查提供一定的理论参考。

研究方法：本研究采用巢式病例对照研究设计，基于2016年1月1日至2017年12月31日在北京市海淀区妇幼保健院建立产检档案、完成常规产检并分娩的妊娠女性所建立的以孕早期为起点的临床妊娠队列，根据血清样本采集孕周一致性（妊娠早期：10-14周；妊娠中期：15-28周；妊娠早期和妊娠中期）将研究人群分为三组，分别将确诊GDM的孕妇作为病例，正常糖耐量（normal glucose tolerance，NGT）人群作为对照，根据孕妇基线年龄（±2岁）进行1:1匹配。采用高通量多因子液相芯片检测技术，定量检测血糖、胰岛素、C-peptide、GIP、PP、sCD40L、Adiponectin、Leptin、Resistin、Adipsin在妊娠女性孕早期及孕中期血清中的水平。构建限制性立方样条函数（restricted cubic splines, RCS）评估妊娠早期及中期血清细胞因子水平与GDM发病风险的非线性关联及暴露-反应关系。使用条件Logistic回归模型来估计孕早期和孕中期空腹血清细胞因子与GDM及其亚型发生风险之间的比值比（odds ratio，OR）及其95%置信区间（confidence interval，CI）。通过Lasso回归和逐步回归法筛选血清细胞因子。结合临床常规危险因素构建预测模型，通过绘制受试者工作特征曲线（receiver operating characteristic curve，ROC curve），计算并比较C-统计量，评估孕早期和孕中期血清各个细胞因子水平对GDM及其亚型发生风险的预测能力。采用Bootstrap法对预测模型进行内部验证。

研究结果：本研究共纳入436名妊娠女性，包括218例GDM病例和218例NGT对照。其中，400名和200名孕妇分别进行了妊娠早期和妊娠中期血清细胞因子测定，184名孕妇同时完成了妊娠早期和妊娠中期检测，基线平均年龄分别为30.18± 2.57岁、30.52±2.50岁和30.43±2.50岁。GDM组的孕前BMI均显著高于NGT对照组（P < 0.05）。在妊娠早期，血清C-peptide、GIP、PP、sCD40L、Resistin及Adipsin水平与GDM发生风险呈显著的线性J型正相关关联（P_{非线性检验} > 0.05），相应细胞因子每增加一个对数单位，GDM的发生风险分别增加138%（OR：2.38，95%CI：1.33-4.40）、100%（OR：2.00，95%CI：1.42-2.85）及43%（OR：1.43，95%CI：1.09-1.90）、96%（OR：1.96，95%CI：1.30-3.03）、79%（OR：1.79，95%CI：1.12-2.92）、494%（OR：5.94，95%CI：2.00-18.50）；妊娠早期血清Leptin水平与GDM发生风险也呈正相关关联（OR：2.29，95%CI：1.57-3.42）；而妊娠早期Adiponectin水平则与GDM发生风险呈显著的线性L型负相关关联（P_{非线性检验} = 0.095），Adiponectin水平每增加一个对数单位，GDM的发生风险降低34%（OR：0.66，95%CI：0.46-0.95）。在妊娠中期，血清C-peptide、GIP及PP水平与GDM发生风险仍然呈仍呈线性的J型关联（P_{非线性检验} > 0.05），其妊娠中期的水平每增加一个对数单位，发生GDM风险的风险分别增加了2.07倍（OR：3.07，95%CI：1.49-6.62)、1.64倍（OR：2.64，95%CI：1.59-4.53)以及0.73倍（OR：1.73，95%CI：1.21-2.53)；妊娠中期Adiponectin水平则与GDM风险仍然显著负相关（OR：0.43，95%CI：0.25-0.72），其余因子与GDM发生风险的关联不显著。妊娠早期至妊娠中期血清C-peptide水平升高或相对升高与GDM风险增大显著相关，而Adiponectin、Resistin和Adipsin水平的升高及相对升高则对GDM风险表现为显著的保护性关联，其他因子的变化及相对变化与GDM风险均无显著关联。结合GDM常规预测因素并加入相应妊娠阶段血清水平差异显著的细胞因子后，妊娠早期和妊娠中期的GDM风险联合预测模型的曲线下面积均为0.85，对不同亚型也均有较好的预测效能。

研究结论：本研究发现，母体妊娠早期血清C-peptide、GIP、PPs、CD40L、Leptin、Resistin及Adipsin水平与GDM发生风险均呈显著正相关关联，而Adiponectin与GDM发生风险呈显著负相关关联；妊娠中期血清C-peptide、GIP及PP水平与GDM发生风险仍然呈显著正相关关联，Adiponectin则与GDM风险呈显著负相关关联，揭示了不同妊娠阶段血清标志物水平对GDM及其亚型风险的差异化调控。此外，结合年龄、BMI等常规GDM预测因子，本研究进一步加入相应妊娠阶段水平差异显著的细胞因子后，妊娠早期和妊娠中期的GDM风险联合预测模型效能较好，为GDM筛查干预窗口前移和早期高危人群识别提供了重要的循证证据和理论依据。

Background: Gestational diabetes mellitus (GDM) has been proved to be associated with an increased risk of both short- and long-term adverse outcomes for mothers and neonates, as well as metabolic abnormalities. Early identification and intervention are pivotal for improving maternal-fetal prognosis. The identification of novel maternal serum biomarkers for risk prediction remains a key challenge in the global prevention and early intervention of GDM. However, previous studies have predominantly focused on a limited number of cytokines or proteins, often within a single gestational window, and have failed to account for the heterogeneity of GDM subtypes based on varying degrees of glucose intolerance. Moreover, inconsistent findings across studies have been reported. To address these gaps, this study systematically selected eight cytokines representing key pathways in glucose and lipid metabolism: serum peptide markers (C-peptide, GIP, and PP), inflammatory cytokine (sCD40L), and adipocytokines (Adiponectin, Leptin, Resistin, and Adipsin). We evaluated the associations between serum biomarkers levels during the first and second trimester and the risk of GDM as well as its subtypes, aiming to construct stage-specific predictive models to facilitate early screening of high-risk populations.

Methods: We conducted a nested case-control study design, using data from a clinical pregnancy cohort established at Haidian Maternal and Child Health Hospital (Beijing, China) from January 2016 to December 2017. Participants were grouped based on serum sample collection timing (early pregnancy: 10-14 weeks; mid-pregnancy: 15-28 weeks; or both). GDM cases were matched in a 1:1 ratio with normal glucose tolerance (NGT) controls by maternal age (±2 years). Serum levels of glucose, insulin, C-peptide, GIP, PP, sCD40L, Adiponectin, Leptin, Resistin, and Adipsin were measured using high-throughput liquid-phase multiplex immunoassays. The non-linear and dose-response associations between serum biomarker levels and GDM risk were assessed using restricted cubic splines (RCS). Conditional logistic regression models were applied to estimate the odds ratios (ORs) and 95% confidence intervals (CIs) for the association between serum biomarker levels and GDM risk. Serum cytokines were selected through a combination of LASSO (Least Absolute Shrinkage and Selection Operator) regression and stepwise regression methods. Prediction models incorporating clinical risk factors and significant biomarkers were constructed and evaluated using receiver operating characteristic (ROC) curves and C-statistics. To assess the internal validity and optimism of the prediction models, we performed bootstrap validation.

Results: A total of 436 pregnant women were included, including 218 GDM cases and 218 NGT controls. Serum biomarkers were measured in 400 and 200 participants during the first and second trimesters, respectively, with 184 participants completing both first and second trimester assessments. The mean baseline age was 30.18±2.57 years for the first trimester group, 30.52±2.50 years for the second trimester group, and 30.43±2.50 years for both trimesters group. Pre-pregnancy BMI was significantly higher in the GDM group compared to the NGT group (P < 0.05). In the first trimester, serum C-peptide, GIP, PP sCD40L, Resistin, and Adipsin levels exhibited a significant J-shaped linear positive association with GDM risk (P _{for non-linear} > 0.05), with an increase of 138% (OR: 2.38, 95%CI: 1.33-4.40), 100% (OR: 2.00, 95%CI: 1.42-2.85), 43% (OR: 1.43, 95%CI: 1.09-1.90), 196% (OR: 1.96, 95%CI: 1.30-3.03), 79% (OR: 1.79, 95%CI: 1.12-2.92), and 494% (OR: 5.94, 95%CI: 2.00-18.50) per log unit increase, respectively. Leptin levels also showed a positive association with GDM risk (OR: 2.29, 95%CI: 1.57-3.42), while Adiponectin was negatively associated with GDM risk, exhibiting an L-shaped linear relationship (OR: 0.66, 95%CI: 0.46-0.95). In the second trimester, C-peptide, GIP, and PP levels remained significantly positively associated with GDM risk, while Adiponectin continued to show a negative association. After incorporating significant biomarkers into the GDM risk prediction models, the areas under the ROC curve (AUC) for the first and second trimesters were 0.85, demonstrating robust predictive ability for different GDM subtypes.

Conclusion: Maternal serum levels of C-peptide, GIP, PP, sCD40L, Leptin, Resistin, and Adipsin during the first trimester are significantly positively associated with GDM risk, while Adiponectin is negatively associated. In the second trimester, C-peptide, GIP, and PP levels remain positively associated with GDM risk, and Adiponectin continues to exhibit a negative correlation. These findings underscore distinct cytokine profiles at different gestational stages differentially modulate GDM risk. Furthermore, by incorporating these biomarkers into GDM risk prediction models, this study provides critical evidence for the advancement of early screening and intervention strategies for GDM, offering significant implications for high-risk population identification and the prevention of GDM.