Prediction of adverse maternal outcomes of early severe preeclampsia

Objectives: Assess soluble FMS-like tyrosine kinase-1/placental growth factor ratio (sFlt-1/PlGF) diagnostic accuracy for predicting adverse maternal outcome in patients with early severe preeclampsia, and whether its predictive performance is superior to full preeclampsia integrated estimate of risk score (PIERS). Study design: Prospective study enrolled patients with early severe preeclampsia (defined by American College of Obstetricians and Gynaecologists 2013 guidelines) admitted to the Clinic for Obstetrics and Gynaecology, Clinical Center of Serbia intensive care unit. Patients underwent delivery to terminate preeclampsia within 48 h of admission. PIERS was generated and blood samples taken at admission. Multiple pregnancies and gestational ages outside 24-34 weeks were excluded. sFlt-1 and PlGF serum concentrations were measured using Elecsys (R) assays and cobas e 601 analyser. Maternal complications were recorded for seven days post-delivery. Main outcome measures: Diagnostic accuracy (sensitivity and specificity), and predictive performance (receiver operating characteristic area under curve [AUC]) vs. PIERS, of sFlt-1/PlGF for predicting adverse maternal outcome. Results: Of 89 patients enrolled, 61 were evaluable. Median frequency of adverse maternal outcomes within seven days of delivery was two. Median sFlt-1/PlGF and PIERS were 521.0 and 5.0%, respectively. sFlt-1/PlGF showed greater correlation with complication number than PIERS (Spearman's rho: 0.728 [p < 0.001] and 0.134 [p = 0.304], respectively). AUC for sFlt-1/PlGF and PIERS were 0.853 and 0.628, respectively. A 377.0 sFlt-1/PlGF cut-off was optimal for predicting complications (75.0% sensitivity; 92.3% specificity). Conclusions: sFlt-1/PlGF correlated more closely with number of adverse maternal outcomes than PIERS, and was a superior predictor of maternal complications.