|
|
|
| Establishing of a model for predicting the risk of preeclampsia based on the "normal" pregnant women |
| 1. Postgraduate Training Base, General Hospital of Northern War Zone of Jinzhou Medical University, Jinzhou, Liaoning Province, 121000； 2. General Hospital of the Northern War Zone, People's Liberation Army, China, Shenyang |
|
|
|
|
Abstract To explore the warning information of pregnant women before their preeclampsia (PE) onset, and to establish a predictive model for the risk of PE occurrence, so as to reduce the incidence of PE. Methods: 176 pregnant women with PE who had undergone regular prenatal checkups in the general hospital of the northern war zone of the Chinese people's liberation army from 2021 to 2022 were collected in study group, and 352 healthy pregnant women with singleton pregnancy who had undergone regular prenatal checkups in hospital during the same period were collected in control group at the ratio of 1:2. The clinical data of the women were recorded and were compared between the two groups. Univariate analysis was used to analyze the influential factors associated with preeclampsia of the women, and the significant influential factors were then selected for the binary logistic regression analysis. A predictive model for preeclampsia was developed, and the area under the curve (AUC) of receiver operator characteristic (ROC) curve and the Hosmer-Lemeshow test were applied to evaluate the efficacy of this model, and the relevant forest diagram was constructed. Results: Binary logistic regression analysis showed that the serum albumin (ALB) level (OR=0.547, 95%CI 0.481-0.622), the body mass index (BMI) value (OR=2.167, 95%CI 1.664-2.821), the urine protein level (OR=2.700, 95%CI 1.448-5.033), the fetal intrauterine growth restriction (FGR) rate (OR=3.030, 95%CI 1.369-6.708) and the edema rate (OR=3.643, 95%CI 1.603-8.281) of the women had significant effects on their preeclampsia occurrence. The predictive model for the risk of the preeclampsia occurrence was established, that is P=1/1+exp(W), and W=-13.911+0.603×serum ALB level (g/L)- 0.773×the change value of BMI (kg/m2) -0.993×urine protein positive situation (1 for the positive urine protein and 0 for the negative or weakly positive urine protein) -1.109×the intrauterine FGR situation (1 for FGR occurrence and 0 for no FGR occurrence) -1.293 ×the edema situation (1 for edema occurrence and 0 for no edema occurrence). The AUC of ROC curve of this predictive model for the preeclampsia occurrence was 0.932 (95%CI 0.906-0.957), and the Hosmer-Lemeshow test performed had showed that there was no statistically significant difference (P>0.05). According to this predictive model, the predictive model-related forest diagram was drawn. Conclusion: The serum ALB level, the change of BMI value, the urine protein level, and the FGR and edema occurrences of the pregnant women are the warning information for their preeclampsia occurrence, and those women who have the warning information should be included in the scope of the key maternal examination. The established prediction model for the risk of preeclampsia occurrence of the pregnant women based on above five indicators can predict the occurrence of preeclampsia in advance, so the achieve early prevention and timely intervention should be conducted to reduce the harm of the preeclampsia of the pregnant women to the mother and child.
|
|
|
|
|
|
|
|
|
