Predictive Models for Preeclampsia in Pregnancy

Preeclampsia remains a leading cause of maternal and perinatal morbidity worldwide. Early prediction is critical to enable targeted surveillance, preventative interventions and resource allocation. Traditional models have relied on maternal demographics and clinical history—such as age, body mass index, parity and prior hypertensive disorders—often combined with simple biophysical measures including mean arterial pressure and uterine artery Doppler indices. Over the past decade, the integration of biochemical biomarkers (for example placental growth factor and pregnancy-associated plasma protein A) has enhanced first-trimester risk stratification. More recently, machine-learning algorithms such as random forest and extreme gradient boosting have been applied to integrate high-dimensional data, demonstrating improved discrimination over classical regression but highlighting the need for rigorous external validation and calibration assessment. Contemporary efforts focus on developing models for early-onset versus late-onset disease, tailoring cut-offs and false-positive rates to clinical utility, and addressing performance in low-resource settings. International collaborations employing individual-participant data meta-analysis have underpinned the creation of new clinical–biochemical–ultrasound multivariable models, while external validation studies have revealed heterogeneity across populations. The global significance of robust prediction lies in optimising antenatal care schedules, guiding prophylactic aspirin use, and facilitating personalised counselling, thereby contributing to reductions in maternal and neonatal complications.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Predictive Models for Preeclampsia in Pregnancy publication trend

The graph below shows the total number of articles in predictive models for preeclampsia in pregnancy across all publications each year (not limited to Nature Index journals).

Technical terms

Area under the receiver operating characteristic curve (AUC): A metric of model discrimination, indicating the probability that a randomly chosen case will have a higher predicted risk than a non-case.

Calibration: The agreement between predicted risks and observed outcomes across the spectrum of risk; poor calibration indicates over- or underestimation of true risk.

Mean arterial pressure (MAP): The average arterial blood pressure over one cardiac cycle, often used as a simple biophysical predictor in first-trimester screening.

Placental growth factor (PlGF): An angiogenic protein produced by the placenta; low maternal serum levels in early pregnancy are associated with increased risk of preeclampsia.

Uterine artery pulsatility index (PI): A Doppler ultrasound measure of vascular resistance in the uterine arteries; elevated PI in the first or second trimester can signal impaired placental perfusion.

References

1. Machine Learning Algorithms Versus Classical Regression Models in Pre-Eclampsia Prediction: A Systematic Review. Current Hypertension Reports (2024).
2. Early Pregnancy Biomarkers in Pre-Eclampsia: A Systematic Review and Meta-Analysis. International Journal of Molecular Sciences (2015).
3. Validation and development of models using clinical, biochemical and ultrasound markers for predicting pre-eclampsia: an individual participant data meta-analysis. Health Technology Assessment (2020).
4. External validation of prognostic models predicting pre-eclampsia: individual participant data meta-analysis. BMC Medicine (2020).