Patient-reported preconceptional characteristics in the prediction of recurrent preeclampsia

To develop a prediction model for recurrent preeclampsia using patient-reported preconceptional characteristics, which can be used for risk stratification of subsequent pregnancies.

Retrospective cohort study using data from The Preeclampsia Registry™ of 1028 women with a history of preeclampsia and at least one subsequent pregnancy.

Candidate predictors were included in a multivariable logistic regression analysis and a backward selection procedure was used to select the final predictors. Internal validation took place by internally validating the model in 500 simulated samples (bootstrapping), which provided a shrinkage factor to create the final model. This final model was evaluated for performance by a calibration plot and the area under the receiver operating curve (AUC). Missing data was handled by multiple imputation.

Recurrent preeclampsia occurred in 467 (45.4%) women. Predictors in the final model were: a history of migraine, first degree relative with cardiovascular disease, first degree relative with placenta-related pregnancy complication, gestational age at delivery of index pregnancy, birthweight of the previous child, history of placental abruption, multiparity, chronic hypertension, interval between index and subsequent pregnancy, paternal non-white ethnicity and maternal age. AUC of the model was 0.63 (95% CI 0.59-0.66). In a subset of women who used aspirin prior or during their subsequent pregnancy, performance of the model was similar (AUC 0.60; 95% CI 0.50-0.71).

In this study we developed a prediction model for recurrent preeclampsia with moderate performance after internal validation. Early risk stratification of subsequent pregnancies that allows for customization of antenatal care and personalized prevention strategies, is not yet possible.