Validation of the fullPIERS model for prediction of adverse outcomes in preeclampsia at a referral center.

Abstract

To validate the use of fullPIERS to predict maternal and perinatal adverse outcomes in a referral center. Cross-sectional study including all pregnant women with preeclampsia (PE) at a referral center in southeast Brazil. The prevalence of PE and adverse outcomes were assessed. The fullPIERS score was tested on three composites of adverse outcomes: maternal adverse outcome; fetal adverse outcomes; and the combination of these two. Furthermore, the fullPIERS risk calculator, was considered to define the cutoff that better estimates adverse outcomes. 2839 women were screened in a one year period, with 208 (7.3%) cases of PE; most were preterm (56.7%); with severe features (74.5%). HELLP syndrome (6.7%), eclampsia (3.8%) and placental abruption (2.4%) were the most frequent complications. FullPIERS assessement had a median of 1.2% (0.45 - 2.3%) and the score had an excelent performance to predict adverse maternal outcome (AUC=0.845, confidence interval 0.776 - 0.914, p-value<0.01). For perinatal adverse outcomes (AUC=0.699, confidence interval 0.581 - 0.816, p-value<0.01) and the composite of maternal and perinatal adverse outcome (AUC=0.804, confidence interval 0.736 - 0.872, p-vale<0.01), fullPIERS score had a suboptimal performance. The cutoff value that best performed for the assessment of maternal adverse outcome was 2.15% (sensitivity of 75% and specificity of 83%). Preeclampsia was a significant complication during pregnancy. The fullPIERS model was an excellent tool to predict maternal adverse outcomes; with a cutoff value of 2.15% in the tested population.