Routine mid-gestational prediction of later preeclampsia

Preeclampsia is thought to be superimposed upon cardiovascular and cardiometabolic risk factors, predominantly consistent with the metabolic syndrome. In this study, we developed and internally validated a prediction model for the development of later preeclampsia in pregnant women at routine second-trimester oral glucose tolerance testing. Data were collected during a prospective clinical cohort study, including pregnant women undergoing routine gestational diabetes mellitus (GDM) screening. Routine clinical data during the GDM screening (e.g., oral glucose tolerance test) were considered as potential predictors. Univariable and multivariable logistic regression with Backward Wald elimination were performed to develop the prediction model. Internal validation was performed using bootstrapping. Predictive performance of the final model was evaluated in terms of discrimination and calibration, both before and after adjusting for overfitting. Of 3227 pregnant women undergoing GDM screening, 137 (4.2 %) subsequently developed preeclampsia. The final prediction model included obstetric history of preeclampsia (yes/no), history of large for gestational age (yes/no), current antihypertensive drug use (yes/no), diastolic blood pressure (mmHg), fasting serum creatinine (μmol/l), fasting serum triglycerides (mmol/l), and urinary protein-creatinine ratio (g/mol creatinine). The area under the receiver operating characteristic curve of the model was 0.79 before and after internal validation, with good model calibration. Upon external validation and impact analysis, the proposed second-trimester preeclampsia prediction model enables accurate estimation of individuals risk on predominantly later third trimester development of preeclampsia. The model could facilitate timely, tailored monitoring and early intervention among pregnant women at risk to improve pregnancy outcomes.