Risk assessment for preeclampsia in nulliparous women at 11–13 weeks gestational age: prospective evaluation of two algorithms

Abstract

Objective To evaluate two algorithms for prediction of preeclampsia in a population of nulliparous women in Norway.Design Prospective screening study.Setting National Centre for Fetal Medicine in Trondheim, Norway.Population Five hundred and forty-one nulliparous women.Methods The women were examined between 11 +0 and 13 +6 weeks with interviews for maternal characteristics and measurements of mean arterial pressure, uterine artery pulsatility index, pregnancy-associated plasma protein A and placental growth factor. The First Trimester Screening Program version 2.8 by The Fetal Medicine Foundation (FMF) was compared with the Preeclampsia Predictor TM version 1 revision 2 by Perkin Elmer (PREDICTOR).Main outcome measures Prediction of preeclampsia requiring delivery before 37 weeks, before 42 weeks and late preeclampsia (delivery after 34 weeks).Results The performance of the two algorithms was similar, but quite poor, for prediction of preeclampsia requiring delivery before 42 weeks with an area under the curve of 0.77 (0.67-0.87) and sensitivity 40% (95% CI 19.1-63.9) at a fixed 10% false positive rate for FMF and 0.74 (0.63-0.84) and sensitivity 30% (95% CI 11.9-54.3) at a fixed 10% false positive rate for PREDICTOR. The FMF algorithm for preeclampsia requiring delivery <37 weeks had an area under the curve of 0.94 (0.86-1.0) and sensitivity of 80% (95% CI 28.4-99.5) at a 10% fixed false positive rate.Conclusions Fetal Medicine Foundation and PREDICTOR algorithms had similar and only modest performance in predicting preeclampsia. The results indicate that the FMF algorithm is suitable for prediction of preterm preeclampsia.

Validation of a first-trimester screening model for pre-eclampsia in an unselected population

“…Regarding external validity, we found conflicting results. Some models performed at a similar level when applied in a different clinical setting from the construction cohort, whereas others underperformed. Oliveira et al .…”

The performance of risk prediction models for pre‐eclampsia using routinely collected maternal characteristics and comparison with models that include specialised tests and with clinical guideline decision rules: a systematic review

“…No external validation of these new models was identified. One additional study externally validated a simple model that used different predictors to the validated models presented here and reported poor performance (model algorithm not reported) …”

“…One additional study externally validated a simple model that used different predictors to the validated models presented here and reported poor performance (model algorithm not reported). 49 The main limitation is that studies did not report model sensitivity and specificity to identify women above the preeclampsia risk level where aspirin is recommended to allow an assessment of clinical value. The majority of studies did not report on model calibration and reported model sensitivity at a fixed specificity of 90 or 95% (corresponding to a 5 or 10% false positive rate).…”

First-Trimester Combined Multimarker Prospective Study for the Detection of Pregnancies at a High Risk of Developing Preeclampsia Using the Fetal Medicine Foundation-Algorithm

“…However, in the first study, the authors did not measure PLGF and in the second one, the authors included in total only 541 nulliparous women [28,29].…”