Department of Public Health, Women's Hospital of Jiangnan University, Wuxi, China.
Center of Reproductive Medicine, Women's Hospital of Jiangnan University, Wuxi, China.
Front Endocrinol (Lausanne). 2023 Sep 29;14:1220472. doi: 10.3389/fendo.2023.1220472. eCollection 2023.
Early diagnosis of gestational diabetes mellitus (GDM) reduces the risk of unfavorable perinatal and maternal consequences. Currently, there are no recognized biomarkers or clinical prediction models for use in clinical practice to diagnosing GDM during early pregnancy. The purpose of this research is to detect the serum G-protein coupled receptor 120 (GPR120) levels during early pregnancy and construct a model for predicting GDM.
This prospective cohort study was implemented at the Women's Hospital of Jiangnan University between November 2019 and November 2022. All clinical indicators were assessed at the Hospital Laboratory. GPR120 expression was measured in white blood cells through quantitative PCR. Thereafter, the least absolute shrinkage and selection operator (LASSO) regression analysis technique was employed for optimizing the selection of the variables, while the multivariate logistic regression technique was implemented for constructing the nomogram model to anticipate the risk of GDM. The calibration curve analysis, area under the receiver operating characteristic curve (AUC) analysis, and the decision curve analysis (DCA) were conducted for assessing the performance of the constructed nomogram.
Herein, we included a total of 250 pregnant women (125 with GDM). The results showed that the GDM group showed significantly higher GPR120 expression levels in their first trimester compared to the normal pregnancy group (p < 0.05). LASSO and multivariate regression analyses were carried out to construct a GDM nomogram during the first trimester. The indicators used in the nomogram included fasting plasma glucose, total cholesterol, lipoproteins, and GPR120 levels. The nomogram exhibited good performance in the training (AUC 0.996, 95% confidence interval [CI] = 0.989-0.999) and validation sets (AUC=0.992) for predicting GDM. The Akaike Information Criterion of the nomogram was 37.961. The nomogram showed a cutoff value of 0.714 (sensitivity = 0.989; specificity = 0.977). The nomogram displayed good calibration and discrimination, while the DCA was conducted for validating the clinical applicability of the nomogram.
The patients in the GDM group showed a high GPR120 expression level during the first trimester. Therefore, GPR120 expression could be used as an effective biomarker for predicting the onset of GDM. The nomogram incorporating GPR120 levels in early pregnancy showed good predictive ability for the onset of GDM.
早期诊断妊娠期糖尿病（GDM）可降低围产期和母体不良结局的风险。目前，尚无公认的生物标志物或临床预测模型可用于临床早期诊断 GDM。本研究旨在检测早孕期血清 G 蛋白偶联受体 120（GPR120）水平，并构建预测 GDM 的模型。
本前瞻性队列研究于 2019 年 11 月至 2022 年 11 月在江南大学附属医院进行。所有临床指标均在医院实验室进行评估。通过定量 PCR 测量白细胞中的 GPR120 表达。然后，采用最小绝对收缩和选择算子（LASSO）回归分析技术对变量进行优化选择，采用多变量逻辑回归技术构建预测 GDM 风险的列线图模型。通过校准曲线分析、受试者工作特征曲线（AUC）下面积（AUC）分析和决策曲线分析（DCA）评估构建的列线图的性能。
本研究共纳入 250 名孕妇（125 名 GDM 患者）。结果显示，GDM 组孕妇在孕早期 GPR120 表达水平明显高于正常妊娠组（p<0.05）。采用 LASSO 和多变量回归分析构建孕早期 GDM 列线图。列线图中的指标包括空腹血糖、总胆固醇、脂蛋白和 GPR120 水平。该列线图在训练集（AUC 0.996，95%置信区间[CI] = 0.989-0.999）和验证集（AUC=0.992）中预测 GDM 的性能良好。列线图的 Akaike 信息准则为 37.961。列线图显示截断值为 0.714（灵敏度=0.989；特异性=0.977）。列线图具有良好的校准和区分能力，同时通过 DCA 验证了列线图的临床适用性。
GDM 组患者在孕早期 GPR120 表达水平较高。因此，GPR120 表达可作为预测 GDM 发病的有效生物标志物。纳入早孕期 GPR120 水平的列线图对 GDM 的发病具有良好的预测能力。
