目的 探讨妊娠期糖尿病（GDM）患者孕24~28周的75 g 口服葡萄糖耐量试验（OGTT）血糖异常项数及妊娠晚期分娩前血糖值与妊娠结局的关系。方法 选择2019年11月—2020年5月在广州医科大学附属妇女儿童医疗中心进行产检并在孕24~28周确诊的167例GDM患者为研究对象,将孕24~28周75 g OGTT结果中仅其中1项时间点血糖异常的孕妇为GDMⅠ组（92例）,2项异常为GDMⅡ组（48例）,3项异常为GDMⅢ组（27例）,比较三组血糖异常项数GDM患者的人口学特点;并分析GDM患者一般人口学特征与妊娠晚期分娩前血糖监测均值的关系,及血糖值对不良妊娠结局的影响。结果 75 g OGTT血糖异常项数与孕前不同的体质指数（BMI）及妊娠晚期的糖化血红蛋白（HbA_1c）间比较差异均有统计学意义（P<0.05）。孕前BMI指数水平对妊娠晚期的空腹血糖、餐后1 h血糖、餐后2 h血糖比较差异均有统计学意义（P<0.05）;75 g OGTT血糖异常项数对空腹血糖及餐后2 h血糖比较差异有统计学意义（P<0.05）;③空腹血糖不同水平组在新生儿低血糖、胎膜早破、早产不良结局中比较,差异有统计学意义（P<0.05）。餐后2 h不同血糖水平间组在新生儿低血糖及胎膜早破中比较,差异有统计学意义（P<0.05）。结论 孕前BMI指数与妊娠中期75 g OGTT的血糖筛查结果有关,75 g OGTT试验中血糖异常项数越多不良妊娠结局的发生概率越大,妊娠期进行规范化的运动饮食干预和必要时的药物干预后可改善妊娠晚期的HbA_1c水平。

Objective To investigate the relationship between abnormal blood glucose items in the 75 g oral glucose tolerance test（OGTT）at 24-28 weeks of pregnancy and the blood glucose levels before delivery in the third trimester of pregnancy and pregnancy outcomes in gestational disbetes mellitus（GDM）patients. Methods All 167 GDM patients diagnosed at 24-28 weeks of gestation in Women and Children's Medical Center Affiliated to Guangzhou Medical University from November 2019 to May 2020 were enrolled as subjects．The pregnant women with only 1 abnormal blood glucose item among the 75 g OGTT results were classified as GDMⅠ group（92 cases）,with 2 abnormal items were GDMⅡ group（48 cases）,and with 3 abnormal items were GDM Ⅲ group（27 cases）．The demographic characteristics of the three groups of GDM patients were compared．The relationship between the general demographic characteristics of GDM patients and the mean value of blood glucose monitoring before delivery in the third trimester was analyzed,and the influence of blood glucose monitoring on adverse pregnancy outcomes was also analyzed．Results ①With different BMI and HbA_1c,there were significant differences in 75 g OGTT blood glucose items（P<0.05）．BMI level had statistically significant effects on fasting blood glucose,1-hour postprandial blood glucose and 2-hour postprandial blood glucose in the third gestational trimester（P<0.05）．②With different number of abnormal blood glucose items,there were significant in fasting blood glucose and 2 hours postprandial blood glucose（P<0.05）．③There were statistically significant differences in the outcomes of neonatal hypoglycemia,premature rupture of membranes and preterm delivery in different fasting blood glucose groups（P<0.05）．There were statistically significant differences in neonatal hypoglycemia and premature rupture of membranes between different 2 hours postprandial blood glucose（P<0.05）. Conclusions BMI can affect the blood glucose screening results of 75g OGTT in the second trimester．The more abnormal blood glucose items in the 75g OGTT test,the greater the probability of adverse pregnancy outcome．Standardized exercise diet intervention and necessary drug intervention during pregnancy can improve the HbA_1c level in the third trimester．

目的 探讨两种不同机器学习算法在妊娠期糖尿病(gestational diabetes mellitus,GDM)风险预测中的应用。方法 选取2019年7月—2020年8月在广州市妇女儿童医疗中心及广东省计划生育专科医院进行产前检查的孕早期妇女520例,其中妊娠期糖尿病孕妇200例,随机抽取同期正常孕妇320例,收集孕妇的一般资料和孕早期(8～12周)的生化指标、血常规和凝血功能等检测资料。利用这些分析变量建立支持向量机(SVM)和Logistic回归(LR)预测模型。根据模型预测能力和模型实用性,如准确率、精确率、真阳性(TP)率、假阳性(FP)率、召回率、F测度、受试者工作特征曲线(ROC)进行效果评价。结果 两种预测模型的分类准确率总体为86%。SVM模型在真阳性(TP)率、假阳性(FP)率、召回率、F测度、受试者工作特征曲线(ROC)方面优于LR模型。结论 在分类与预测方面,支持向量机算法比Logistic回归模型更具有实用价值。

Objective To explore the application of two different machine learning algorithms in the risk prediction of gestational diabetes mellitus (GDM). Methods A total of 520 pregnant women with gestational diabetes mellitus were selected from Women and Children's Medical Center and Guangdong Family Planning Hospital from July 2019 to August 2020, including 200 cases of gestational diabetes mellitus, and 320 normal pregnant women in the same period. The general information of pregnant women and the detection data of biochemical indexes, blood routine test and coagulation function in early pregnancy (8～12 weeks) were collected. Support vector machine (SVM) and logistic regression (LR) prediction models were established by using these analysis variables. According to the predictive ability and practicability of the model, something like accuracy rate, precision ratio, true positive (TP) rate, false positive (FP) rate, recall rate, F-measure and receiver operating characteristic curve (ROC) were evaluated. Results The classification accuracy of the two models was 86%. SVM model is better than LR model in TPrate, FPrate, recall rate, F measure and ROC. Conclusion Support vector machine is more practical than logistic regression model in classification and prediction.