孕早期铁的营养状况与妊娠糖尿病发生的关联

The relationship between iron nutritional status in early pregnancy and the occurrence of gestational diabetes

目的探讨孕早期铁的营养状况与妊娠糖尿病(GDM)发生的关联。方法选择2019年2月至2021年1月在北京市朝阳区妇幼保健院产科建档立卡的814例孕妇作为研究对象,根据孕妇是否发生GDM分为GDM组(n=79)和非GDM组(n=735),收集一般资料并通过调查问卷及半定量食物频数表收集孕妇孕早期生活方式和饮食情况。通过拟合曲线分析孕早期铁的营养状况与GDM发生概率的关系,采用多因素Logistic回归分析GDM发生的影响因素,构建贝叶斯网络模型,使用Netica软件进行贝叶斯网络推理,并对模型效能进行评价。结果糖尿病家族史、孕前BMI、膳食血红素铁、膳食总铁是导致GDM发生的独立危险因素(P<0.05)。拟合曲线显示,当膳食血红素铁>3.55mg/d和膳食总铁>21.09mg/d时,GDM发生概率呈上升趋势。糖尿病家族史、孕前BMI、膳食血红素铁、膳食总铁通过网络关系与GDM建立联系,糖尿病家族史、孕前BMI、膳食血红素铁、膳食总铁与GDM的发生均有直接联系。构建的贝叶斯网络模型的一致性指数(C-index)为0.841(95%CI:0.829~0.863),ROC曲线下面积(AUC)为0.850(95%CI:0.833~0.874),具有较高的区分度,校准曲线表明模型整体预测的准确性较高。结论糖尿病家族史、孕前BMI、膳食血红素铁、膳食总铁是导致GDM发生的独立危险因素。当膳食血红素铁>3.55 mg/d和膳食总铁>21.09 mg/d时,GDM发生概率呈上升趋势,适当控制孕早期铁摄入量,可有效降低妊娠糖尿病的发生。

Objective To explore the relationship between iron nutritional status in early pregnancy and gestational diabetes(GDM).Methods From February 2019 to January 2021,814 pregnant women who were registered in the obstetrics department of Beijing Chaoyang Maternal and Child Health Hospital were selected as the study subjects,and the lifestyle and diet of pregnant women in the early pregnancy were collected through questionnaires and semi quantitative food frequency tables.According to whether pregnant women have GDM,they are divided into GDM group(n=79)and non GDM group(n=735).Multivariate Logistic regression was used to analyze the influencing factors of GDM.The relationship between iron nutritional status in early pregnancy and the probability of GDM was analyzed by fitting curve.Build Bayesian network model,use netica software for Bayesian network reasoning,and evaluate the effectiveness of the model.Results Family history of diabetes,pre pregnancy BMI,dietary heme iron and total dietary iron were independent risk factors for GDM(P<0.05).The fitting curve showed that the probability of GDM increased with the increase of dietary heme iron and total dietary iron(P<0.05).Family history of diabetes,pre pregnancy BMI,dietary heme iron,and total dietary iron are linked to GDM through network relationships.Family history of diabetes,pre pregnancy BMI,dietary heme iron,and total dietary iron are directly related to the occurrence of GDM.The consistency index(C-index)of the constructed Bayesian network model is 0.841(95%CI:0.829–0.863),and the area under the ROC curve(AUC)is 0.850(95%CI:0.833–0.874),which has high discrimination.The calibration curve shows that the overall prediction accuracy of the model is high.Conclusion Family history of diabetes,pre pregnancy BMI,dietary heme iron and total dietary iron are independent risk factors for GDM.When dietary heme iron>3.55 mg/d and total dietary iron>21.09 mg/d,the probability of GDM increased.Appropriate control of iron intake in early pregnancy can effectively reduce the incidence of gestational diabetes.