First trimester prediction of early onset preeclampsia using demographic, clinical, and sonographic data: a cohort study

Objective The aim of this research was to evaluate the performance of a predictive model for early onset preeclampsia (PE) during early gestation. Method Prospective multicenter cohort study was performed in women attending 11-14 weeks ultrasound. Medical history and biometrical variables were recorded and uterine artery Doppler was performed. All patients were followed until postpartum period. Constructed predictive models were compared using the area under the associated receiver operating characteristic curve. Sensitivity, specificity, and likelihood ratios were estimated for each outcome. Results A total of 627 patients were enrolled. Sixty-five (10.4%) developed gestational hypertension, of which 29 developed PE (4.6% of the total sample) and nine occurred before 34 weeks (1.5% of total sample). Prediction model generated for early onset PE (ePE) with 5% false positive achieve sensitivity of 62.5% and specificity of 95.5%. The positive and negative likelihood ratios for ePE were 13.9 and 0.39, respectively. Development of ePE was significantly associated with history of preterm labor (p = 0.002) and diabetes mellitus (p = 0.02). Conclusions This study confirms the advantage of combining multiple variables for prediction of ePE. (C) 2013 John Wiley & Sons, Ltd.