New prognostic prediction tool for suspected early-onset pre-eclampsia: model development and performance evaluation.

Abstract

Objectives (1) To develop a risk score for early-onset pre-eclampsia leading to delivery within one week from repeated marker determinations on pregnancies with an sFlt-1/PlGF ratio above 38, and (2) to compare it (i) with sFlt-1/PlGF ratio model and (ii) with the sFlt-1/PlGF ratio 655 cutoff. Design Retrospective cohort study. Setting Oviedo, Spain. Sample 522 blood samples from 363 singleton pregnancies with clinical suspicion of pre-eclampsia between 27 and 34 weeks of gestation. Methods NT-proBNP, sFlt-1 and PlGF were combined using linear mixed models with random intercept based on 213 samples from 123 pregnancies with an sFlt-1/PlGF ratio above 38. Main outcome measures Early-onset pre-eclampsia diagnosis leading to delivery within one week from assessment. Results None of the 253 pregnancies with an sFlt-1/PlGF ratio of 38 or below developed early-onset pre-eclampsia. The prognostic prediction tool included sFlt-1 MoM, NT-proBNP and gestational age at time (GA) of repeated measurements. The area under the ROC curve (AUC) for early-onset pre-eclampsia diagnosis leading to delivery within one week was 88.247 (95% CI 0.822-0.934) for the prediction tool and 82.639 (95% CI 0.752-0.892) for the sFlt-1/PlGF + GA model (P=0.04). At an sFlt-1/PlGF ratio 655 cutoff the detection ratio was 31.9% (19.1-47.1) with false positive rate of 4.2% (1.7-8.5). With the same false positive rate, the detection rate with the prognostic prediction tool was 53.2% (38.1-67.9) (P=0.03). Conclusions A prediction tool derived from NT-proBNP, sFlt-1 MoM and gestational age linear mixed model provided clinically useful prediction of early-onset pre-eclampsia prognosis when clinically suspected.