妊娠早期全血细胞计数联合孕妇一般特征对妊娠期糖尿病的综合预测能力探讨

First-trimester complete blood count combined with maternal characteristics as a predictor of gestational diabetes mellitus

目的探讨妊娠早期全血细胞计数联合孕妇一般特征对妊娠期糖尿病(GDM)的综合预测作用。方法收集2015年5月至2018年7月于复旦大学附属上海市第五人民医院进行规律产检的孕妇共1412例,其中GDM患者258例,正常健康孕妇1154人。记录所有孕妇的临床资料及妊娠早期全血细胞计数。通过logistic回归分析GDM相关风险因素,确立GDM预测模型,比较各模型对GDM的预测能力。结果Logistic回归分析显示,年龄、妊娠前体重指数、既往GDM史、糖尿病家族史、中性粒细胞和淋巴细胞比值以及白细胞、中性粒细胞和单核细胞计数均为GDM发生的独立预测因素。在所有孕妇中,中性粒细胞和单核细胞联合纳入一般特征模型后,预测GDM发生的受试者工作特征曲线下面积(AUC-ROC)为0.809[综合判别改善指数(IDI)=0.056,P=0.001],优于其他模型(一般特征模型,AUC-ROC=0.753;单核细胞+一般特征模型,AUC-ROC=0.764;中性粒细胞+一般特征模型,AUC-ROC=0.775)。在所有孕妇中筛选出无既往GDM史孕妇,重复上述分析得到相似的结果。结论在一般特征模型基础上,联合纳入妊娠早期中性粒细胞计数和单核细胞计数可以进一步提高对GDM发生的预测能力。

Objective To Investigate comprehensive predictive ability of first-trimester complete blood count combined with maternal characteristics for gestational diabetes mellitus(GDM).Methods From May 2015 to July 2018,1412 pregnant women were retrospectively screened at the Fifth People′s Hospital of Shanghai,Fudan University.We recruited 258 women who developed GDM and 1154 women who had normal glucose level during pregnancy.At the first visit,clinical data and complete blood count result were obtained.GDM prediction models were established through logistic regression analysis of GDM related risk factors and the prediction abilities of each model were compared.Results Logistic regression analyses identified age,pre-pregnancy body mass index,previous GDM history,family history of diabetes mellitus,the neutrophil-to-lymphocyte ratio,leukocyte,neutrophil,and monocyte counts were significantly independent predictors of GDM.In the entire cohort,the predictive ability of neutrophil and monocyte counts together with maternal basal characteristics model for the development of GDM[areas under the receiver operating characteristic curve(AUC-ROC)=0.809,integrated discrimination improvement(IDI)=0.056,P=0.001]was the best among various models(basal characteristics model,AUC-ROC=0.753;Monocyte count+basal characteristics model,AUC-ROC=0.764;neutrophil count+basal characteristics model,AUC-ROC=0.775).Similar results obtained by the same way in all pregnant women without previous GDM history.Conclusion It could improve the prediction of GDM with model incorporated maternal characteristics and first-trimester neutrophil and monocyte counts.