Factors influencing recurrence of preeclampsia in pregnant women with a history of preeclampsia and the establishment of a predictive model

Introduction

Preeclampsia (Pre-eclampsia), also known as preeclampsia, is a pregnancy-specific hypertensive disease, which usually occurs after 20 weeks of pregnancy, characterized by elevated blood pressure and proteinuria, which may be accompanied by damage to the function of other maternal organs [1], 2]. Unlike eclampsia, preeclampsia is the state of preeclampsia, which may develop into eclampsia without proper treatment. Recurrent preeclampsia is a serious complication of pregnancy, which is omni-directional and poses a serious threat to pregnant women and fetuses [3], [4], [5]. For pregnant women, recurrent preeclampsia may lead to cardiovascular system damage and increase the risk of cardiovascular disease; at the same time, it may lead to renal dysfunction and even renal failure; it may also cause liver function damage, abnormal blood coagulation and systemic edema [6]. These complications not only affect the health of pregnant women, but also lead to dangerous conditions during delivery, such as heart failure, pulmonary edema and so on [7], [8], [9], [10]. For the fetus, recurrent preeclampsia may lead to intrauterine hypoxia, affecting its normal growth and development. Due to impaired placental function, the fetus may be at risk of growth restriction, resulting in a lower body weight than normal at the same gestational age. In addition, recurrent preeclampsia may also increase the risk of premature delivery or stillbirth, posing a serious threat to the life and safety of the fetus [11]. In addition to direct hazards, recurrent preeclampsia may also have a long-term impact on the future health of pregnant women and fetuses [12], [13], [14]. For example, pregnant women may face a long-term risk of cardiovascular disease after preeclampsia, while fetuses affected by hypoxia and malnutrition may suffer irreversible damage to their nervous system, affecting their future cognitive and motor function. Previous studies by Wainstock et al. [15]. have shown that cesarean section and preterm labor are important risk factors for recurrent preeclampsia, but there is a lack of relevant research on model prediction that combines factors. The establishment of predictive models also aids in strengthening the self-management of pregnant women. By understanding the results of predictive models, pregnant women can gain a clearer understanding of their condition and risks, thereby more actively participating in prenatal care and disease prevention and enhancing their self-care awareness and capabilities [16], [17], [18]. Therefore, this study analyzes the factors influencing the recurrence of PE in pregnant women with a history of the condition and further establishes a predictive model to provide a reference for clinical practice.

Subjects and methods

Results

Development of a predictive model

Based on the results of the logistic regression analysis, age, gestational age at onset during the previous pregnancy, interpregnancy interval, and blood pressures from the previous pregnancy were incorporated into a model predicting the recurrence of preeclampsia in pregnant women with a history of the condition. The model formula is Logit(P)=(0.910Age)+(0.987Age of onset of previous pregnancy)+(1.167Gestational interval)+(1.186Systolic blood pressure in previous pregnancy)+(0.970Diastolic blood pressure in previous pregnancy). The clinical benefits are shown in Figure 1. The results of DCA analysis showed that the model had good clinical practicability. The model curve demonstrates superior performance by being positioned closer to the upper right corner compared to the curve representing all interventions (ALL), suggesting that interventions guided by the model yield greater benefits than blanket interventions across the board, as illustrated in Figure 1. To construct the calibration curve, the prediction probabilities are categorized into distinct buckets, employing either a ‘uniform’ or ‘quantile’ division strategy. Within each bucket, the average prediction probability of all samples is computed and designated as the abscissa (x-axis). Concurrently, the probability of positive instances within the same bucket is calculated and assigned as the ordinate (y-axis). By sequentially connecting these calculated points, the calibration curve is effectively plotted, providing a visual representation of the model’s predictive accuracy. This model showed calibration curves with slopes close to 1 in both the training and validation sets, indicating good consistency between the predicted risks and the actual risks, as shown in Figures 2 and 3.

Figure 1:

Predictive model. Note: Successful establishment of prediction model.

Figure 2:

Training set calibration curve. Note: The slope of the calibration curve is close to 1, indicating that the risk is in good agreement with the actual risk.

Figure 3:

Validation set calibration curve. Note: The slope of the calibration curve is close to 1, indicating that the risk is in good agreement with the actual risk.

ROC curve analysis of the predictive model’s value in forecasting recurrence of preeclampsia in pregnant women with a history of preeclampsia

ROC analysis demonstrated that in the training set, the AUC for predicting the recurrence of preeclampsia in pregnant women with a history of the condition was 0.898 with a standard error of 0.034 as shown in Figure 4. In the validation set, the AUC was 0.991 with a standard error of 0.001 as shown in Figure 5.

Figure 4:

Training set roc curve. Note: The area under the ROC curve is approximately one, indicating that the prediction value is good.

Figure 5:

Validation set ROC curve. Note: The area under the ROC curve is approximately one, indicating that the prediction value is good.

Discussion

The risk of recurrence of preeclampsia increases significantly during second pregnancy, which poses a great threat to maternal and infant health. Therefore, to explore the influencing factors of its recurrence and establish an effective prediction model is of great significance to improve the quality of pregnancy management and reduce the risk of maternal and infant complications [19], 20]. In this study, by comparing the general data of patients in the verification group and the model group, and the results of this study, to explore the influencing factors of recurrence, and based on these factors to build a predictive model, and then verify.

The results showed that there were statistically significant differences between patients in the Recurrence group and Non-recurrence group in age, gestational week starting during the previous pregnancy, pregnancy interval, systolic blood pressure during the previous pregnancy, and systolic blood pressure during the previous pregnancy (p<0.05). Multivariate logistic results further demonstrated that these factors are independent influencing factors on recurrence. One possible explanation for these findings is that as women age, their physiological functions gradually decline and the elasticity of their blood vessels decreases, potentially leading to reduced adaptability to hypertension during pregnancy [21], [22], [23]. Additionally, aging is associated with an increased risk of various chronic diseases, which may further impact blood pressure control during pregnancy [24], 25]. Earlier onset of disease in a previous pregnancy may indicate more severe impairment of endothelial function in pregnant women, or the presence of some yet unidentified genetic factors or pathophysiological processes [26], [27], [28]. These factors could be reactivated during a subsequent pregnancy, leading to an earlier or more severe onset of preeclampsia. A shorter inter-pregnancy interval (less than 5 years) may affect a woman’s physical recovery and adaptability. A short interval might mean the body has not fully recovered before becoming pregnant again, with previous pregnancy blood pressure levels directly reflecting the woman’s blood pressure condition [29], [30], [31]. Higher blood pressure levels may indicate some degree of vascular dysfunction or damage, which could persist or worsen in a subsequent pregnancy, leading to the recurrence of preeclampsia.

This study constructed a mode (Logit(P)=(0.910Age)+(0.987Age of onset of previous pregnancy)+(1.167Gestational interval)+(1.186Systolic blood pressure in previous pregnancy)+(0.970Diastolic blood pressure in previous pregnancy) l to predict the risk of recurrent PE in pregnant women with a history of PE.The model included age, gestational age, gestational interval, systolic blood pressure and diastolic blood pressure of previous pregnancy as predictive factors.The results of the study show that women with advanced age, shorter gestational weeks after the onset of previous pregnancies, shorter pregnancy intervals, and higher systolic and diastolic blood pressures in previous pregnancies are at greater risk. The model built in this study inputs data such as age, age of onset of first pregnancy, pregnancy interval, systolic and diastolic blood pressure of previous pregnancies, and the algorithm can calculate the probability of recurrence risk. For example, a 35-year-old pregnant woman with a first pregnancy of 28 years, a pregnancy interval of 2 years, and a previous pregnancy with a systolic blood pressure of 150 mmHg and a diastolic blood pressure of 100 mmHg, put the data into the model. The result is that at 337, the risk threshold indicates that she has a high risk of recurrent PE. Doctors can add preventive measures in advance to reduce the probability of recurrence.The performance evaluation of the model shows that the slope of the calibration curve is close to 1 in both the training set and the verification set, which strongly indicates that there is a good consistency between the risk predicted by the model and the actual observed risk. In the training set, the results of ROC analysis showed that AUC of the model for predicting recurrent preeclampsia was 0.898, and the standard error was 0.034 When the optimal truncation value is set to 0.85, the sensitivity of the model is 93.4 % and the specificity is 91.1 %, indicating that the model has high accuracy in identifying pregnant women who will actually have preeclampsia again. At the same time, the proportion of women who mistakenly identify low-risk pregnant women as high-risk is also relatively low. In the verification set, the performance of the model is further verified. ROC analysis results show that the AUC is as high as 0.991 and the standard error is 0.001. When the optimal truncation value is 0.94, the sensitivity of the model is 93.5 %, and the specificity is 100.0 %. The prediction model based on Logistic regression shows good accuracy and reliability in predicting the risk of recurrent PE in pregnant women with a history of preeclampsia. It provides clinicians and pregnant women with a powerful tool to assess and manage the risk of second pregnancy.

Limitation

Although this study has made some achievements in predicting the risk of This time the researchers, there are still some shortcomings. First, the size and diversity of the sample could be limiting factors. Although an adequate number of samples were collected as much as possible, a larger sample size might provide more representative and stable results. Additionally, the diversity of the sample may affect the generalizability of the model, that is, the model’s predictive effectiveness across different populations. Furthermore, the validation and evaluation of the model also face challenges. While this study used training and validation sets to assess the model’s performance, more external datasets and real-world application scenarios might reveal limitations and deficiencies of the model in practical use. Future research should aim to expand the sample size, consider more potential influencing factors, and explore more advanced predictive models and methods to enhance prediction accuracy and stability.

Conclusions

To sum up, this study provides new ideas and tools for predicting, preventing and controlling the recurrence of preeclampsia in pregnant women with a history of preeclampsia. Through further research and practice, the predictive model is combined with clinical practice to give full play to its value.