Prediction model of preeclampsia risk and construction method and application thereof
By using pre-pregnancy body mass index and adjustment variables for restrictive cubic spline analysis in the pre-eclampsia risk prediction model, the problem of inaccurate pre-eclampsia risk prediction in the prior art was solved, and high-accuracy risk prediction and risk threshold determination were achieved.
Patent Information
- Application Number
- CN202510204228.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-24
- Publication Date
- 2025-06-20
AI Technical Summary
The prior art is difficult to accurately predict the risk of preeclampsia, resulting in inaccurate prediction results.
Using pre-pregnancy body mass index as the input factor, ethnicity, assisted reproduction, and whether or not a primacy are adjusted variables, restricted cubic spline analysis was performed to obtain the correlation curve between pre-pregnancy body mass index and pre-eclampsia risk, determine the risk threshold, and verify the accuracy of the prediction results through the correlation between weight gain during pregnancy.
Accurate prediction of preeclampsia risks is achieved, the prediction accuracy is improved, and the risk threshold is determined, providing a reliable basis for judging preeclampsia risks.
Smart Images
Figure CN120183686A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of biomedicine, and particularly relates to a prediction model for preeclampsia risk, a construction method thereof, and an application thereof. Background Art
[0002] Preeclampsia is a multi-system progressive disease unique to pregnancy, with new-onset hypertension and proteinuria after 20 weeks of gestation, or without proteinuria but accompanied by end-organ dysfunction. Preeclampsia is the main cause of increased risk of maternal and child death and long-term chronic diseases. As a complex disease caused by multiple factors, its pathogenesis is not yet clear, resulting in inaccurate prediction results of the preeclampsia risk prediction model. Therefore, a preeclampsia risk prediction model with high accuracy is needed. Summary of the Invention
[0003] To solve the above problems, the present invention provides a prediction model for preeclampsia risk, a construction method thereof, and an application thereof.
[0004] A prediction model for preeclampsia risk, wherein the prediction model takes the prepregnancy body mass index as an input factor, takes ethnicity, assisted reproduction, and whether it is a primipara as adjustment variables, and performs restricted cubic spline analysis to obtain the curve of the correlation between the prepregnancy body mass index and preeclampsia risk;
[0005] The result of the prediction model is that the correlation between BMI and preeclampsia risk presents an inverted L-shaped curve, and the inflection point of the curve of the correlation between the prepregnancy body mass index and preeclampsia risk is used as the risk threshold to evaluate preeclampsia risk.
[0006] Preferably, the risk threshold is 21.5 kg / m 2 .
[0007] Preferably, restricted cubic spline analysis is performed with the gestational weight gain as an input factor to obtain the result of predicting the gestational weight gain, and the result of predicting the gestational weight gain is used to judge the accuracy of the prediction result of the prediction model.
[0008] Preferably, the correlation between the gestational weight gain and the risk of preeclampsia presents a J-shaped curve, and the low-risk interval is 10.94 kg to 15.90 kg. The result of predicting the gestational weight gain is obtained according to the low-risk interval;
[0009] The prediction result of the prediction model is consistent with the result of predicting the gestational weight gain, and the accuracy of determining the risk of preeclampsia is 100%.
[0010] The construction method of the prediction model includes the following steps:
[0011] Collect clinical data, where the clinical data includes predictors and preeclampsia conditions;
[0012] Perform a chi-square test using the said predictors, and select the predictors with a P < 0.05 correlation between pregnant women and preeclampsia risk as the first correlation predictors;
[0013] Perform a logistic regression analysis using the said first correlation predictors, select the first correlation predictor with the largest OR value as the second correlation predictor, and select external factor adjustment variables from the remaining said first correlation predictors;
[0014] Perform a restricted cubic spline analysis using the said second correlation predictor and the adjustment variables to obtain a prediction model, and judge the preeclampsia risk according to the risk threshold in the prediction model;
[0015] The said second correlation predictor is the prepregnancy body mass index;
[0016] The said adjustment variables are ethnicity, assisted reproduction, and whether the woman is a primipara.
[0017] Preferably, the said predictors are ethnicity, age, height, prepregnancy weight, whether the woman is a primipara, whether there is assisted reproduction, weight before delivery, mode of delivery, gender of the newborn, weight of the newborn, and one-minute Apgar score.
[0018] Preferably, the said first predictors are ethnicity, whether there is assisted reproduction, whether the woman is a primipara, prepregnancy body mass index, and pregnancy weight gain.
[0019] The application of the said prediction model in improving the accuracy of predicting preeclampsia risk.
[0020] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0021] The prediction model of the present invention takes the prepregnancy body mass index as the input factor, and takes ethnicity, assisted reproduction, and whether the woman is a primipara as the adjustment variables, and determines that the correlation between the prepregnancy body mass index and the preeclampsia risk presents an inverse L-shaped curve, and the inflection point is 21.5 kg / m 2 , and determines the risk threshold, providing an accurate basis for judging the preeclampsia risk.
[0022] The present invention uses univariate and multivariate logistic regression analyses to determine the influence of different factors on the risk of preeclampsia; uses a logistic regression model with restricted cubic splines to test the relationship between the prepregnancy body mass index and pregnancy weight gain and preeclampsia. The restricted cubic splines show that the correlation between the prepregnancy body mass index and the preeclampsia risk presents an inverse L-shaped curve, and the inflection point is 21.5 kg / m 2;The association between pregnancy weight gain and the risk of preeclampsia shows a J-shaped curve. When pregnancy weight gain is between 10.94 kg and 15.90 kg, the risk of preeclampsia is the lowest. Description of the Drawings
[0023] Figure 1 Forest plot of multivariate logistic regression analysis.
[0024] Figure 2 It is a restricted cubic spline graph of variables and the occurrence of preeclampsia. A is the pre-pregnancy body mass index, B is the weight gain during pregnancy, and C is the restricted cubic spline graph of the weight gain during pregnancy and the occurrence of preeclampsia in different groups of pre-pregnancy body mass index (underweight, normal weight, overweight, and obesity). Detailed Implementation Modes
[0025] The following is a detailed description of the specific implementation modes of the present invention. However, it should be understood that the protection scope of the present invention is not limited by the specific implementation modes. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention. The experimental methods described in the embodiments of the present invention are all conventional methods unless otherwise specified.
[0026] Embodiment
[0027] 1. General Information
[0028] Pregnant women who had regular antenatal care and established files in Urumqi Youai Hospital from January 2020 to June 2024 were collected as the research objects, and were divided into the preeclampsia group and the non-preeclampsia group according to whether preeclampsia occurred.
[0029] Inclusion criteria: ① The diagnosis of preeclampsia refers to the "Diagnosis and Treatment Guidelines for Hypertensive Disorders Complicating Pregnancy (2020)"; ② Singleton pregnancy; ③ Age 18 - 49 years old; ④ No other pregnancy complications and comorbidities.
[0030] Exclusion criteria: ① History of neurological diseases, infectious diseases, immune diseases, and cardiovascular diseases such as hypertension and hyperlipidemia, and history of diseases such as kidney disease before pregnancy; ② Those with incomplete basic information.
[0031] Record information such as their names, hospital numbers, ethnic groups, ages, diagnoses, heights, pre-pregnancy weights, whether they are primiparas, whether they have assisted reproduction, weights before delivery, delivery methods, genders of newborns, weights of newborns, one-minute Apgar scores, and contact numbers.
[0032] This invention has been approved by the Ethics Committee of Urumqi Youai Hospital (approval number: WLMQYALL2022003), and all participants have given informed consent.
[0033] 2. Pre-pregnancy body mass index and pregnancy weight gain
[0034] Based on the height and weight of the pregnant woman before pregnancy and the weight at the time of approaching delivery, calculate the pre-pregnancy body mass index (BMI) and pregnancy weight gain (GWG) according to the formula.
[0035] BMI = self-reported pre-pregnancy weight (kg) ÷ square of the height (m) measured at the first antenatal care.
[0036] GWG = weight at delivery - pre-pregnancy weight.
[0037] According to the recommended value standard for weight gain in pregnant women during pregnancy (WS / T 801-2022), the pre-pregnancy body mass index is classified into underweight [BMI: <18.5 kg / m 2 , normal weight [BMI: 18.5 kg / m 2 ~23.9 kg / m 2 , overweight [BMI: 24 kg / m 2 ~27.9 kg / m 2 , and obesity [BMI: ≥28 kg / m 2 .
[0038] National Health Commission of the People's Republic of China. Recommended value standard for weight gain in pregnant women during pregnancy: WS / T 801-2022 [S]. 2022.
[0039] 3. Statistical analysis
[0040] Use JMP 14.0 and R version 4.4.0 for statistical analysis. Data conforming to the normal distribution are expressed as mean ± standard deviation and the independent samples t-test is used for two independent sample data; data not conforming to the normal distribution are expressed as median and quartiles (P 25 , P 75 ), and the non-parametric Mann-Whitney U test is used for comparison of two independent samples; count data are described by the number of cases and the constituent ratio (n, %), and the x 2 test is used between groups; a logistic regression model with restricted cubic splines is used to test the non-linear relationship between pre-pregnancy body mass index and pregnancy weight gain and preeclampsia. P < 0.05 is considered statistically significant.
[0041] Results
[0042] 1. Comparison of baseline data and pregnancy outcomes
[0043] A total of 13,294 pregnant women were included in the study, including 559 pregnant women with preeclampsia (4.20%), and 12,735 pregnant women in the non-preeclampsia group (90.63%). The incidence of preeclampsia in underweight women was 1.72%, in normal weight women was 2.85%, in overweight women was 6.60%, and in obese women was 16.05%. There were statistically significant differences in the ethnicity, prepregnancy weight, prepregnancy body mass index, whether they were primiparas, whether they used assisted reproductive technology, weight gain, delivery method, one-minute Apgar score, and neonatal birth weight among pregnant women in different groups (P<0.05), as shown in Table 1.
[0044] Table 1 Comparison of baseline data and pregnancy outcomes of pregnant women in different groups
[0045]
[0046] 2. Multivariate logistic regression of preeclampsia
[0047] Regarding the occurrence of preeclampsia as an event, the factors with P<0.05 in the above univariate analysis and those that may affect the occurrence of preeclampsia (prepregnancy weight, prepregnancy body mass index, prepregnancy weight gain, ethnicity, whether they were primiparas, whether they used assisted reproductive technology) were used as independent variables and included in the logistic regression analysis. The logistic regression equation was output using the forward conditional method to obtain the OR and the 95% confidence interval (CI) of OR, and a forest plot was drawn. The results are shown in Table 2 and Figure 1 。
[0048] The output logistic regression equation is:
[0049]
[0050] Taking the example of all variables being included in the model, let X1 be ethnicity (coded as in the univariate part above), X2 be whether assisted reproductive technology was used, X3 be primipara, X4 be prepregnancy body mass index (coded as in the univariate part above), X5 be prepregnancy weight, X6 be weight gain, and β0=-7.45.
[0051] It should be noted that the coefficients here are the coefficients in the multivariate analysis. For categorical variables such as ethnicity variables, they should be substituted and calculated according to their specific coding methods. For example, if it is ethnic group B, X1 = 1, and ethnic group A, X1 = 0, etc.
[0052] Results showed that for each one-unit increase in pre-pregnancy body mass index, the risk of preeclampsia increased by 1.04%; for each one-unit increase in body weight, the risk of preeclampsia increased by 1.07%; the likelihood of preeclampsia in group C pregnant women was 2.14 times that of group A pregnant women, OR 95% CI (1.55 - 2.96); the likelihood of preeclampsia in women using assisted reproductive technology was 2.12 times that of women not using assisted reproductive technology, OR 95% CI (1.33 - 3.37); the likelihood of preeclampsia in primiparas was 1.92 times that of non-primiparas, OR 95% CI (1.55 - 2.36); the likelihood of preeclampsia in overweight pregnant women was 1.68 times that of normal-weight pregnant women, OR 95% CI (1.30 - 2.18); the likelihood of preeclampsia in obese pregnant women was 3.16 times that of normal-weight pregnant women, OR 95% CI (2.08 - 4.79). There was a significant positive correlation between pre-pregnancy body mass index, weight gain, group C, primiparity, and cesarean section and the occurrence of preeclampsia. See Table 2 for details. Figure 1 。
[0053] Table 2 Results of univariate and multivariate logistic regression
[0054]
[0055] The present invention shows that there are significant differences in the incidence of preeclampsia among different ethnic groups. In the univariate analysis, the risk of preeclampsia in ethnic minorities is significantly increased, and the risk is the highest in ethnic group C (OR = 2.42). In the multivariate analysis, the risk of ethnic minorities is still significant, but the effect value is weakened, indicating that after controlling other variables, the risk of these ethnic groups is still relatively high, which may be related to various factors such as genetics, lifestyle, eating habits, and socioeconomic status. The univariate analysis shows that the risk of preeclampsia in women using assisted reproductive technology is significantly increased (OR = 2.26), and this risk is still significant in the multivariate analysis (OR = 2.12), indicating that assisted reproductive technology may be an independent risk factor for preeclampsia. The possible mechanisms include epigenetic aberrations leading to placental abnormalities, the absence of luteal secretion factors, and the immune response to allogeneic gametes. The risk of preeclampsia in primiparous women is significantly increased in both univariate and multivariate analyses (OR = 1.70 and 1.92), which may be related to the adaptability of primiparous women to pregnancy changes, placental formation, and maternal vascular adaptability changes. The proportion of cesarean sections in the preeclampsia group is significantly higher than that in the non-preeclampsia group, which may be because preeclampsia increases the risks of the mother and fetus, and doctors may be more inclined to choose cesarean section to reduce complications during childbirth. The birth weight and Apgar score in the preeclampsia group are both lower than those in the non-preeclampsia group. The median and interquartile range of the birth weight in the preeclampsia group are lower than those in the control group, 1540 (960g, 1920g) vs 3135 (2850g, 3440g), indicating that preeclampsia may have a negative impact on fetal growth and development and the health status after birth.
[0056] In the present invention, underweight before pregnancy shows a lower risk of preeclampsia in the multivariate analysis (OR = 0.81), while overweight and obesity significantly increase the risk of preeclampsia, especially in the obese group (OR = 3.16). Both pre-pregnancy weight and gestational weight gain are positively correlated with the risk of preeclampsia. For every 1-unit increase in pre-pregnancy weight, the risk of preeclampsia increases by 4% (OR = 1.04), and for every 1-unit increase in gestational weight gain, the risk increases by 7% (OR = 1.07), which may be related to the metabolic and vascular adaptability changes caused by weight gain. Overweight / obesity is considered a chronic inflammatory disease, which can increase the levels of plasma C-reactive protein and certain inflammatory cytokines. The metabolic and biochemical disorders associated with overweight and obesity lead to an increase in neutrophils, releasing toxic compounds (i.e., reactive oxygen species and myeloperoxidase), which can attack and damage the integrity of vascular endothelial cells. This mechanism ultimately leads to the clinical symptoms of preeclampsia. This emphasizes the importance of pre-pregnancy weight management in preventing preeclampsia.
[0057] 3. Logistic regression restricted cubic spline analysis of pre-pregnancy body mass index and gestational weight gain in relation to preeclampsia
[0058] Using restricted cubic spline regression analysis, with 4 knots, the model adjusted for ethnicity, assisted reproduction, and parity.
[0059] The correlation between pre-pregnancy body mass index and the risk of preeclampsia showed an inverse L-shaped curve, with an inflection point at 21.5 kg / m 2 ; when the pre-pregnancy body mass index exceeded this value, for each additional unit increase in BMI, the risk of preeclampsia increased significantly ( Figure 2 A).
[0060] The association between gestational weight gain and the risk of preeclampsia showed a J-shaped curve. The risk of preeclampsia was lowest when GWG was between 10.94 - 15.90 Kg ( Figure 2 B). Discussing the relationship between GWG and preeclampsia by grouping pre-pregnancy body mass index, it was found that when the BMI of pre-pregnancy underweight pregnant women exceeded 21.63 kg and the BMI of pre-pregnancy normal weight pregnant women exceeded 10.94 kg, the risk of preeclampsia increased significantly ( Figure 2 C).
[0061] The prediction results of gestational weight gain and the risk of preeclampsia were consistent with those of pre-pregnancy body mass index and the risk of preeclampsia, and the accuracy of determining the risk of preeclampsia was 100%.
[0062] This invention shows that the relationship between pre-pregnancy body mass index and the risk of preeclampsia is not a simple linear increase or decrease, but rather that after a certain specific BMI value, the risk rises significantly. This may mean that when the pre-pregnancy body mass index is below 21.5, the pregnant woman's physical condition has a stronger protective effect against preeclampsia, or the risk factors for preeclampsia have not been significantly activated. After exceeding this value, for each additional unit increase in BMI, the risk of preeclampsia increases significantly, which may be related to obesity-related physiological changes such as inflammation, insulin resistance, endothelial dysfunction, etc. These changes may promote the development of preeclampsia. It can be recommended that women with a pre-pregnancy body mass index exceeding 21.5 take more active weight management and health intervention measures to reduce the risk of preeclampsia.
[0063] When the GWG is between 10.94 - 15.90 Kg, the risk of preeclampsia is the lowest. This may mean that within this GWG range, the nutritional status of pregnant women and fetuses is better balanced, without either undernutrition or excessive weight gain, thus reducing the risk of preeclampsia. Beyond this range, especially with excessive GWG, the risk of preeclampsia may increase. This may be related to the physiological stress caused by excessive weight gain. For example, increased adipose tissue may release more inflammatory factors, affecting vascular function and thus increasing the risk of preeclampsia. These results emphasize the importance of pre-pregnancy body mass index and gestational weight management in preventing preeclampsia. The weight control of pregnant women before and during pregnancy should be regarded as an important strategy to reduce the risk of preeclampsia.
[0064] When conducting a subgroup analysis of gestational weight gain in pregnant women with different pre-pregnancy body mass indexes, it was found that when the BMI of pre-pregnancy underweight pregnant women exceeded 21.63 kg and the BMI of pre-pregnancy normal weight pregnant women exceeded 10.94 kg, the risk of preeclampsia increased significantly. A study in Reunion Island with a 18.5-year follow-up involving 57,000 singleton pregnant women showed an independent association between optimal pregnancy weight gain and late-onset preeclampsia, reducing the risk of late-onset preeclampsia. The adjusted odds ratio (OR) was 0.74, P = 0.004.
[0065] In summary, preeclampsia is the result of the combined action of multiple factors. It is necessary to strengthen the publicity of preeclampsia prevention and control knowledge, advocate a healthy lifestyle and eating habits, conduct trials on lifestyle interventions before pregnancy, focus on screening high-risk groups, and take weight loss measures as early as possible to effectively slow down or prevent the occurrence and development of preeclampsia and improve the pregnancy outcomes of pregnant women.
[0066] It should be noted that when the claims of the present invention involve numerical ranges, it should be understood that any value between the two endpoints of each numerical range and the two endpoints can be selected. To avoid repetition, the present invention describes preferred embodiments.
[0067] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they know the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present invention.
[0068] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these changes and modifications.
Claims
1. A prediction model for the risk of preeclampsia, characterized in that: The prediction model uses the pre-pregnancy body mass index as an input factor, and uses ethnicity, assisted reproduction, and whether or not a primipara is an adjustment variable, and performs restricted cubic spline analysis to obtain a curve of the correlation between the pre-pregnancy body mass index and the risk of preeclampsia; The curve of the correlation between the pre-pregnancy body mass index and the risk of preeclampsia is an inverted L-shaped curve, and the inflection point of the curve of the correlation between the pre-pregnancy body mass index and the risk of preeclampsia is used as the risk threshold to assess the risk of preeclampsia.
2. The prediction model according to claim 1, characterized in that The risk threshold is 21.5 kg / m 2 .
3. The prediction model according to claim 1, characterized in that The restricted cubic spline analysis is performed with the gestational weight gain as an input factor to obtain the result of the gestational weight gain prediction, and the result of the gestational weight gain prediction is used to judge the accuracy of the prediction result of the prediction model.
4. The prediction model according to claim 3, characterized in that The correlation between weight gain during pregnancy and the risk of preeclampsia presents a J-shaped curve, with a low-risk interval of 10.94 kg to 15.90 kg, and the result of the weight gain prediction during pregnancy is determined based on the low-risk interval; The prediction results of the prediction model are consistent with the results of the pregnancy weight gain prediction, and the accuracy of determining the risk of preeclampsia is 100%.
5. The method for constructing a prediction model according to claim 1, characterized in that: The following steps are involved: Collect clinical data including predictors and status of preeclampsia; The chi-square test was performed using the predictors, and the predictor with a correlation P < 0.05 between pregnant women and the risk of preeclampsia was selected as the first correlation predictor; Using the first correlation prediction factors to perform logistic regression analysis, selecting the first correlation prediction factor with the largest OR value as the second correlation prediction factor, and selecting external factors from the remaining first correlation prediction factors as adjustment variables; Performing restricted cubic spline analysis using the second correlation predictor and the adjustment variable to obtain a prediction model, and determining the risk of preeclampsia according to a risk threshold in the prediction model; The second correlation predictor is pre-pregnancy body mass index; The adjustment variables are ethnicity, assisted reproduction, and whether or not the mother is a primipara.
6. The construction method according to claim 5, characterized in that: The predictive factors are ethnicity, age, height, pre-pregnancy weight, whether or not a primipara, whether or not assisted reproduction, pre-delivery weight, mode of delivery, newborn sex, newborn weight, and one-minute Apagar score.
7. The construction method according to claim 5, characterized in that: The first predictive factors are ethnicity, whether assisted reproduction is used, whether the mother is a primipara, pre-pregnancy body mass index, and weight gain during pregnancy.
8. Use of the prediction model described in claim 1 in improving the accuracy of predicting the risk of preeclampsia.