Prospective evaluation of the Simple Fetal Medicine Foundation's first‐trimester prediction model for preeclampsia in twin pregnancies

Introduction There is no widely applicable prediction model for preeclampsia (PE) in twins, the existing models lack strict external validation. Therefore, this study aimed to evaluate the predictive performance of the Simple Fetal Medicine Foundation model (i.e., Simple FMF model, only including the maternal risk factors and mean arterial pressure [MAP]) for the prediction of PE at 11–13 + 6 weeks' gestation in twin pregnancies. Material and Methods This was a prospective, non‐intervention study in 464 twin pregnancies (including dichorionic twins and monochorionic twins) at 11–13 + 6 weeks' gestation in a tertiary center between September 1, 2021 and October 31, 2024. Maternal characteristics, medical, obstetric, and drug history were recorded. MAP was measured according to regular protocol. The FMF's algorithm is freely available online, which was used for the calculation of patient‐specific risk of PE in each participant. The performance of screening for PE occurring at different gestational weeks by the Simple FMF model was evaluated with the adjustment of aspirin use. The efficiency of the screening strategy with maternal factors and MAP was evaluated using the receiver operating characteristic curve, and the detection performance for PE at a fixed false‐positive rate was also analyzed. Results Among the 464 women included, 24.14% (112/464) developed PE, among which 1.29% (6/464) occurred before 30 weeks, 2.80% (13/464) occurred before 32 weeks, 6.47% (30/464) occurred before 34 weeks, 19.18% (89/464) occurred before 37 weeks, 4.96% (23/464) (including 21 cases of dichorionic diamniotic twins and 2 cases of monochorionic diamniotic twins) occurred at 37 weeks or later. The Simple FMF model had area under the receiver operating characteristic curves of 0.911 (95% confidence interval [CI]: 0.830–0.992), 0.799 (95% CI: 0.668–0.930), 0.797 (95% CI: 0.712–0.882), and 0.767 (95% CI: 0.706–0.828) for the screening of PE occurring before 30, 32, 34, and 37 weeks, respectively. Conclusions This study has demonstrated that the Simple FMF model was effective in predicting PE in twin pregnancies, especially for cases of PE occurring before 30, 32, 34, and 37 weeks. This approach only uses maternal characteristics, medical history, and MAP, making it suitable for implementation in resource‐limited settings.