Development and validation of prediction model for pre‐eclampsia‐related maternal adverse outcome: retrospective cohort study

ABSTRACT Objectives Pre‐eclampsia (PE) is a major contributor to maternal morbidity and mortality. This study aimed to identify clinical risk factors for maternal adverse outcome in women with PE, and to assess the time to event using predictive modeling. Methods This retrospective cohort study included women with a singleton pregnancy diagnosed with PE who received antenatal care between January 2019 and December 2023 at one of three hospitals in China. Participants were assigned randomly into a training set and a validation set in a 7:3 ratio. Clinical characteristics and laboratory parameters collected at the time of PE diagnosis were analyzed to identify risk factors for maternal adverse outcome using univariate and multivariate logistic regression. Cox regression analysis was employed to examine factors associated with time to maternal adverse outcome. Two predictive nomogram models were developed: Model 1, which was based on multivariate logistic regression and predicts the absolute risk of maternal adverse outcome; and Model 2, which was derived from Cox regression coefficients and estimates the likelihood of maternal event‐free survival at 3, 5 and 7 days after PE diagnosis. Model performance was assessed using the area under the receiver‐operating‐characteristics curve (AUC), calibration curves and decision‐curve analysis. Results A total of 1520 women diagnosed with PE were screened, of whom 1400 were included based on predefined inclusion and exclusion criteria; 979 were assigned to the training set and 421 to the validation set. Multivariate logistic regression identified the following variables as significant independent risk factors for maternal adverse outcome: maternal age ≥ 35 years, prepregnancy body mass index (underweight, overweight or obese), irregular antenatal care, earlier gestational age at presentation, visual disturbance, mean arterial pressure ≥ 120 mmHg, elevated 24‐h urine protein, elevated soluble fms‐like tyrosine kinase‐1 to placental growth factor ratio, low hemoglobin, elevated aspartate aminotransferase, total bilirubin > 15.9 μmol/L and serum urea > 7.5 mmol/L. Cox regression confirmed that these factors were associated significantly with time to maternal adverse outcome, with the exception of elevated total bilirubin and serum urea. Model 1 demonstrated excellent predictive performance, with an AUC of 0.93 (95% CI, 0.92–0.95) in the training set and 0.92 (95% CI, 0.89–0.95) in the validation set. Model 2 achieved AUC values of 0.81 (95% CI, 0.75–0.87), 0.86 (95% CI, 0.82–0.89) and 0.89 (95% CI, 0.87–0.92) in the training set and 0.79 (95% CI, 0.70–0.88), 0.82 (95% CI, 0.75–0.89) and 0.88 (95% CI, 0.84–0.92) in the validation set for predicting maternal adverse outcome at 3, 5 and 7 days after PE diagnosis, respectively. For each model, calibration curves showed strong agreement between predicted and observed probabilities, while decision‐curve analysis confirmed the clinical utility of both models. Conclusions This study identified key clinical predictors of maternal adverse outcome in women with PE and developed two nomogram models for risk stratification. The validated predictive models offer reliable tools for the early identification of high‐risk patients, facilitating timely clinical intervention to improve maternal outcome. © 2025 International Society of Ultrasound in Obstetrics and Gynecology.