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

王芳,孙田歌,李悦,黄新梅,吴跃跃,于志艳,盛励,陈灶萍,张瑞,臧淑妃,丁和远,查兵兵,刘军
DOI: https://doi.org/10.3760/cma.j.cn311282-20210219-00103
2021-01-01
Abstract: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, 1 412 pregnant women were retrospectively screened at the Fifth People′s Hospital of Shanghai, Fudan University. We recruited 258 women who developed GDM and 1 154 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.
What problem does this paper attempt to address?