关键词: 先兆子痫, 危险因素, 列线图, 模型，统计学, 高龄孕妇

Abstract:

Objective: To explore the risk factors for pre-eclampsia(PE) complicating elderly pregnant women, and construct and validate a nomogram risk prediction model. Methods: The clinical data of 1 287 elderly pregnant women admitted to the Suqian First Hospital, Jiangsu Province from January 2017 to August 2024 were retrospectively analyzed. The pregnant women were randomly divided into a training set (n=858) and a validation set (n=429) at a ratio of 2∶1. According to the pregnant women in the training set were further divided into a complication group (n=94) and a non-complication group (n=764). The clinical data of the two groups were compared. Binary logistic regression analysis was used to screen the influence factors for PE complicating elderly pregnant women, and a nomogram model was constructed based on the results. The Bootstrap method was used for internal validation of the model. The concordance index (C-index) was calculated, and calibration curve, receiver operating characteristic (ROC) curve, and decision curve were drawn to evaluate the predictive performance of the model. The validation set data were used for further internal validation of the model. Results: Binary logistic regression analysis showed that a high pre-pregnancy body mass index (BMI), a family history of hypertension, a previous history of PE, a history of spontaneous abortion, gestational diabetes mellitus, high serum uric acid, and malnutrition were all risk factors for PE complicating elderly pregnant women (all P<0.05), while a high platelet count and regular prenatal check-ups were protective factors (both P<0.05). A nomogram risk prediction model for PE complicating elderly pregnant women was constructed based on the results of the multivariate analysis. After internal validation, the C-indices of the training set and the validation set were 0.841 and 0.823 respectively, and the calibration curves were closed to the ideal curves. The ROC curves showed that in the training set, the sensitivity, specificity, area under the curve (AUC), and optimal cut-off value were 87.10%, 81.88%, 0.859 and 316 points respectively; in the validation set, they were 83.87%, 82.61%, 0.842 and 310 points respectively. Decision curve analysis showed that when the threshold probability of the training set were between 0.03-0.74 and 0.77-0.82, and that of the validation set were between 0.02-0.70, 0.73-0.76 and 0.83-1.00, a higher net benefit could be obtained. Conclusions: Pre-pregnancy BMI, family history of hypertension, previous history of PE, history of spontaneous abortion, gestational diabetes mellitus, platelet count, serum uric acid, malnutrition, and regular prenatal check-ups are all influencing factors for PE complicating elderly pregnant women. The nomogram risk prediction model for PE complicating elderly pregnant women constructed based on these factors has high predictive performance and can guide the early clinical screening of high-risk pregnant women for timely intervention.

Key words: Pre-eclampsia, Risk factors, Nomograms, Models, statistical, Elderly pregnant women