Objective: To assess the value of a prediction model for pre-eclampsia (PE) in the first trimester (Ultrasound Obstet Gynecol 2007;30:742-794) for the prediction of late (>34 weeks) and early (< or =34 weeks) PE in a high-risk population.
Methods: Longitudinal study performed in 152 high-risk pregnancies with at least one high-risk condition: previous PE, hypertension, pregestational diabetes, renal disease, obesity, hyperlipidemia, autoimmune disorders, thrombophilia or recurrent pregnancy loss. Mean uterine artery pulsatility index at 11 to 13 + 6 weeks and a series of maternal variables were combined in order to obtain the estimated 'a posteriori risk for PE' in each woman. This risk for unaffected women was compared with that for patients who subsequently developed late and early PE. The performance of such approach was described by receiver-operating characteristic curves.
Results: Late PE developed in 13 (8.6%) pregnancies and early PE in seven (4.6%). The median 'a posteriori risk for PE' in the unaffected, late PE, and early PE groups was 0.62%, 1.22%, and 2.49% (P < 0.01), respectively. For a false-positive rate of 10%, the detection rates of late and early PE were 23.1 and 42.9%, respectively.
Conclusions: This referenced model shows a modest performance when applied to high-risk women.