The value of fasting plasma glucose and lipid profiles between 7 and 15 gestational weeks in the prediction of gestational diabetes mellitus

Objective: To explore the value of using fasting plasma glucose (FPG) and lipid profiles between 7 and 15 gestational weeks to predict gestational diabetes mellitus (GDM). Methods: The medical records of 2 138 pregnant women who had prenatal care in Beijing Obstetrics and Gynecology Hospital from August 2011 to February 2012 were analyzed retrospectively. According to results of the oral glucose tolerance tests, women were devided into the GDM group (n=240) and the normal group (n=1 898). Maternal characteristics, FPG and lipid levels between 7 and 15 gestational weeks were compared between the two groups. Logistic regression analysis and receiver operator characteristics(ROC) curve were used in the analysis. Results: Potential markers for the prediction of GDM included total cholesterol, triglyceride (TG) , low-density lipoprotein cholesterol/high-density lipoprotein cholesterol ratios (LDL-C/HDL-C) , triglyceride to high-density lipoprotein cholesterol ratios (TG/HDL-C) and FPG. After adjustment of confounding factors, age (OR=1.046, 95%CI:1.003-1.090), pre- pregnancy body mass index (OR=1.104, 95%CI: 1.049-1.161), gravidity>3 (OR=1.768, 95%CI:1.071-2.920), FPG (OR=8.137, 95%CI:5.412-12.236), TG (OR=1.460, 95%CI:1.148-1.858) were independently associated with the risk of developing GDM. Equation, PGDM=1/{1+exp[-(-16.542+0.045×age+0.103×pre-pregnancy body mass index+0.551×gravidity>3+2.110×FPG+0.372×TG)]}, was constructed by the logistic regression analysis. Sensitivity (67.5%) and specificity (70.5%) were determined by the calculated risk score, with a cut-off value of 0.11 (area under the curve: 0.751, 95%CI:0.718-0.783, P<0.001). Conclusions: FPG and TG, together with clinical characteristics may have a better predictive value for the risk of GDM.