A preliminary study of serum metabolic markers in the early prediction and diagnosis of gestational diabetes mellitus
Objective To identify serum metabolic markers for early prediction and diagnosis of gestational diabetes mellitus(GDM).Methods A retrospective case-control study was conducted.The study subjects were from pregnant women enrolled in the Birth Cohort Study of the Women's Hospital,Zhejiang University,from1 November 2018 to 30 March 2020.100 cases of GDM(GDM group,Age 36.03±3.91)and 150 non-GDM pregnant women matched for clinical information(control group,Age35.49±3.46)were retrospectively selected for the study.Fasting serum samples were collected at 15-20 weeks of gestation(prior to GDM diagnosis,T1 period)and 24-28 weeks of gestation(during GDM diagnosis,T2 period).Liquid chromatography-tandem mass spectrometry(LC-MS/MS)was used to quantify GDM-related serum metabolic small molecules,including 1,5-anhydroglucitol,3-hydroxybutyric acid,phenylalanine,and isoleucine.These molecules,along with basic clinical information(age,gestational week,BMI)and standard biochemical indicators(FPG),were used to develop predictive models for the early detection of GDM at T1 and the diagnosis of GDM at T2.Statistical analysis was performed using t-tests or Mann-Whitney U-tests.Result The results of the targeted quantitative validation study indicate:At the T1 stage,the level of 1,5-anhydroglucitol was found to be significantly lower(P=0.001)in the GDM group compared to the control group.Conversely,the level of isoleucine was significantly higher(P=0.027)in the GDM group.There were no significant differences in the levels of 3-hydroxybutyrate and phenylalanine between the two groups(P＞0.05).The combination of the 4 metabolites yielded the highest predictive value(AUC)for GDM at T1,with an AUC of 0.670(95％CI:0.602-0.739),P＜0.001.At the T2 stage,the GDM group had significantly lower levels of 1,5-anhydroglucitol(P＜0.05)and significantly higher levels of 3-hydroxybutyric acid and isoleucine(P＜0.05)than the control group,with no significant differences in phenylalanine levels(P=0.626).The combination of the four metabolites had the highest diagnostic value(AUC)for GDM,0.717(95％CI 0.651-0.783),P＜0.001.The analysis of seven different combinations of GDM prediction/diagnostic models created by combining four metabolites with basic clinical information and routine biochemical indicators showed:We found that the AUC value of the GDM diagnostic model built with FPG,BMI,pre-pregnancy BMI,age,gestational week,and the 4 metabolite indicators in T2 stage was the best,0.794(95％CI 0.736-0.851),P＜0.001,with a sensitivity of 72％;The best AUC value for the GDM prediction model built with the same indicators at T1 was 0.711(95％CI 0.646-0.776),P＜0.001,with a sensitivity of 77％.Conclusions Four metabolic small molecules,1,5-anhydroxyglucitol,3-hydroxybutyric acid,phenylalanine,and isoleucine,were integrated with clinical indicators(FPG)and clinical information(age,gestational week,BMI)to develop a predictive model for GDM at gestation(T1)and a diagnostic model for GDM at gestation(T2),demonstrating promising clinical prediction and diagnostic capabilities.1,5-Anhydroglucitol,3-hydroxybutyric acid,phenylalanine,and isoleucine show potential as valuable markers for the prediction and diagnosis of GDM.
万方数据
维普
